Synthesis and Properties of Diorganomagnesium Compounds

University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange

Doctoral Dissertations Graduate School

12-1967

Conrad William Kamienski University of Tennessee - Knoxville

Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss

Part of the Chemistry Commons

Recommended Citation Kamienski, Conrad William, "Synthesis and Properties of Diorganomagnesium Compounds. " PhD diss., University of Tennessee, 1967. https://trace.tennessee.edu/utk_graddiss/3077

This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council:

I am submitting herewith a dissertation written by Conrad William Kamienski entitled "Synthesis and Properties of Diorganomagnesium Compounds." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the equirr ements for the degree of Doctor of Philosophy, with a major in Chemistry.

Jerome F. Eastham, Major Professor

We have read this dissertation and recommend its acceptance:

Bruce M. Anderson, George K. Schwitzer, David A. Shirley

Accepted for the Council:

Carolyn R. Hodges

Vice Provost and Dean of the Graduate School

(Original signatures are on file with official studentecor r ds.) Novemb er 27, 1967

To the Graduate Council:

I am submitting herewith a dissertation written by Conrad Wiliiam Kamienski entitled "Synthesis and Properties of Diorgano magnesium Compounds." I recommend that it be accepted in partial fulfillment of the requirements for the degree of Doctor of Philosophy, with a maj or in Chemistry .

We have read this dissertation and recommend its acceptance:

Accepted for the Council:

Vice President for Graduate Studies and Research SYNTHESIS AND PROPERTIES OF DIORGANOMAGNESIUM COMPOUNDS

A Dissertation

Presented to

the Graduate Council of

The University of Tennessee

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Conrad William Kamienski

December 1967 To my dear wife , Diane

769836 ACKNOWLEDGEMENT

The author wishes to acknowledge the kindness and generosity, as well as the guidance , shown him by Profeasor J. F. Eastham during his studies at the University of Tennessee and during the preparation of this thesis.

Many thanks are also due to the Lithium Corporation of America and to the kindness , understanding and foresight of its president , Mr.

Harry D. Feltenstein as well as the help of many other of its employees, without whose support this achievement could not have been possible.

Gratitude is also due to the author' s father, Mr. Walter

Warchalowski , and to other members of his family , for the wonderful guidance and understanding shown him throughout his career.

A spec ial note of thanks mu st also be give� to Professor M. z. von Krzywoblocki of Michigan State University, wh ose kindness and advice inspired the author to pursue this course of study .

iii TABLE OF CONTENTS

CHAPTER PAGE

I. INTRODUCTION • • • • • • • • • • • • • • . • • • • • • • 1

A. Methods of Preparation of Dialkylmagnesium Compounds • • 1

B. Properties of Unsolvated Dialkylmagnesium Compounds 8

C. Solvation and the Properties of Dialkylmagnesium

Compounds ...... 12

D. Relative Chemical Reactivity of Magnesium Alkyls and

Grignard Reagents • • • • • • • • • • • • • • 18

E. Metallation and Polymerization by Magnesium Alkyls 2 6

II . RESULTS AND DISCUSSION • • • • • • • • • • • • 30

•• A. Preparation of Dialkylmagnesium Reagents . 30

1. Preparation of R2Mg by Metal-Metal Exchange

between RLi and RMgCl or MgCl2 • • • • • 30

2. Direct Preparation of R2Mg from Magnesium

Metal and Alkyl Chlorides • • • • • • • . 40

B. Physical Properties of Dialkylmagnesium Reagents 44

C. Complexes of Dialkylmagnesium Reagents with

Bridgehead Diamines ••. •••••• 57

D. Me tallations with Dialkylmagnesium Reagents 60

E. Reactions of Dialkylmagnesium Reagents with

Benzophenone • • • • • • • 78

F. Miscellaneous Reactions with Dialkylmagnesium Reagents • 85

iv ... ..

CHAP TER PAGE

II . (CONTINUED)

1. Attemp ted Addition of Dialkylmagnesium

Reagents to Ethylene • • • • • • • • . . . • • • • • 85

2. Attempted Reaction of di-,!.-Butylmagnesium

Reagents with Phenyldimethylchlorosilane

in Ethy l Ether, • • • • • • • • • • · • . . .. . 85

3. Effect of Dialkylmagnesium Reagents on the

Halogen-Metal Interconversion Reaction with

Alkyllithium Reagents ...... 86

G. Homologization 'of Alkyllithium Reagents • • • • • • • 81

1. Two Ordinary Runs ...... ' 88

2. Four TMEDA-Catalyzed Runs ...... 95

3. S-Butyllithium Run • ...... • i . . . 100

4. Summary ...... 103

: III . EXPERIMENTAL ••• ...... 107

A .. Introductory Information . • . . . . • ...... • . 107

1. Reagents . • ...... • • • . . . . • • 107

. 2. Instrumentation and Analytical Techniques • . . '108

B. Preparation of Dialkyl�gnesium Reagents from

Grignard and Organolithium Reagents . . . • • . 110

1. Preparation of Di-S-butylmagnesium . • • . . . . • 110

2. Preparation of Methylisobutylmagnesium.• • • • 111

3. Preparation of Methyl-,!-butylmagnesium • . • . . .. 112

4. Preparation of S-Butylphenylmagnesium .· • • • • • • 112 vi

·CHAPTER ·PAGE

III. (CONTINUED)

5. Preparation of Diisobutylmagnesium • . . . . 113

6. Preparation of Di-�-butylmagnesium from

Simultaneous Preparation of.n-Butyllithium -

and-�-Butylmagnesium Chloride ...... 114

C. Preparation of Dialkylmagnesium Reagents from

Organolithium Reagents and Anhydrous Magnesium

Chloride • • • • ...... 114

I 1. Preparation of Di-�-butylmagnesium ...... 114

a. From dimethyl ether-solvated by-product

MgC12 ...... 114

b. From ether-desolvated �y-product MgC 12 . .. 115 c. From ether-desolvated MgC 12 by-product

from reaction of a Grignard reagent

with chlorine ...... • . . . . . 116

d. From dime thyl ether-solvated commercial

anhydrous MgCl 2 ...... • 117

e. From ethyl ether-ac tivated, desolvated

commercial anhydrous MgCl2 . • • • • . . . . 118

f. From unsolvated by-product MgC12 •.••• , • 118

g. From unsolvated ball-milled lump commercial

anhydrous MgC.l2 • • • • • • • • • • • • • . • 119

h. From alcohol-activated, desolvated commercial

anhydrous MgC12 • • . • • • • •.• • • • • • • 120

i. From desolvated commercial MgCl2"6H20 • • • • 121

j. From commercial powdered anhydrous MgCl 2 • . . 122 .. vii

CHAPTER PAGE.

III . · (CONTINUED)

• • 2. Preparatien ef Eli-S-butylmagnesium • . • • • • • 122

a. From dimethy l ether-solvated by-product ...... •· ...... • • ' 122

b. From ether-desolvated MgC12 by�product from

reac tion of a Grignard reagent with benzyl

chloride ...... 123

3. Preparation of .Di-,!!-butylmagnesium. . . . . 124 '

4. Preparation of .�ineopentyl magnesium • . .. 125

D. Direct Preparation of Dialkylmagnesium. Reagents

From Magnesium Metal and Alkyl Chlorides • ...... 126·

• • • 1. Preparation of ·:Di-,!-butylmagnesium • . . . 1.26

a. In dimethyl ether-cyclohexane solution • • 126 . b. By vacuum-stripping of solvent from the

GrignaTd reagent , followed by re-solution

. in benzene •• ••••. ••...... 127

c. In diethyl ether-benzene solution . • ...... 127

• • 2. Preparation of Di-,!!�.lntylmagnesium • • • • • • 128

• • 3. Preparation of Di•,!!•8Dylmagnesium • • • • e-1 •.J .. 129

4. • Preparation of »iisopropylmagnesium . • • . • . e1 • •• 130

E. PrepaTation of Mixed Soluble Di-,n-�tylmaanes ium.

Complexes with ,!·Butyllithium and. a-Butyl lithium 131

1. Preparation of '»1-,!!·butyl•aneaium • • • • 131

2. Preparation of Butyllithium-Butylmaanesium

Complexes ...... 132 viii

CHAPTER P"AGE

III. (CONTINUED)

F. Preparation of Soluble , Mixed Dialkylmagnesium

Complexes from Insoluble Dialkylmagnesium

Compounds ...... •. 133

1. Preparation of a Soluble Diisobutyl -ni-L-butyl-

magnesium Complex ...... 133

2 . Preparation of a Soluble Diisobutyl-Di-

!-�tylmagnesium Complex • • • • • • • 133

3. Preparation of a Soluble D1-n•butyl•Di -.

��butylmagnesium Complex • • • • • • • 133

· G. Insoluble Complexes of Dialkylmagnesium Reagents

with Triethylenediamine and ·

Methyltriethyl�nediamine • • • ...... 134

1. Complexes with Triethylenediamine ...... 134

a. Di-L-butylmagnesium- triethylenediamine

complex ...... 134

b. Other dialkylmagnesium- triethylenediamine

complexes ...... 135

2 . Complexes with Methyltriethylenediamine . . . . . 136

a. Di -.!,.but¥lmag nesium-methyltriethylendiamine

complex 136 .

b. Other dialkylmagnesium-methyltriethylene-

diamine complexes ...... 136

H. Metalations Effected with Dialkylmagnes ium Reagents 137 ix

CHAPTER PAGE

III. {CONTINUED)

1. Me tallation of Resorcinol Dimethyl Ether 137

a. In the ab sence of base . 137

b. In the presence of base 138

2. Me tallation of Thiophene ..• 139

a. Me tallation without Lewis base 139

b. Attempted metallation with Lewis base 140

c. Me tallation with THF plus potassium

t-butoxide • • 140

d. Me tallation in the presence of �-butyl lithium

plus TMEDA • 141

3 . Me tallation of n-Picoline 141

4. Metallation of Toluene • • 142

a. Attempted metallation 142

b. Me tallation with partially pyrolyzed

di-�-butylmagnes ium 143

5. Metallat ion of Benzene with Partially Pyrolyzed

Di-�-butylmagnesium 144

6. Me tallation of Alkynes • 145

a. Acetylene 145

b. 3 -Methyl-1-butyne 146

c. 4-Me thyl-1-pentyne • ...... 146

7. Me tallation of Dimethyl Sul foxide 147

I. Attempted Hy drogenolysis of Dialkylmagnesium

Reagents • • • • • • . 148 X

CHAPTER PAGE

III. (CONT INUED)

1. With and Without Added TMEDA • • • • • • • • • 148

2 . In .the Presence of Potassium S-Butoxide Alone

and witn Added THF and TMEDA • • • • • • • • 148

3. In the Presence of �-Butyllithium and a Limited

Amount of TMEDA 149

J. Attempted Addition of Dialkylmagnesium Reagents

to Ethylene ...... 149

K. Reactions of Dialkylmagnes ium Reagents with

Ketones (Benzophenone) • • • • • • • 150

1. Di-S-butylmagnes ium and Benzophenone • 1.50

2. Di-�-butylmagnesium and Benzophenone • . . . . 151

a. NMR studies on the colored complexes . . . . 151

b. Study of the products of reaction . . . . 153

. L. Attempted Reaction of Dialkylmagnesium Reagent with

Pheny ldimethylchlorosilane • . • • • 154

M. Effect of Dialkylmagnesium Reagents on the Halogen-

Metal Interconversion Reaction with Alkyllithium

Reagents • • ...... 154

N. Homologization of Alkyllithium Reagents and

Dimethyl Ether • • . . 156

1. �-Butyllithium ••••• . . . . . 156

a •. Determination of products of reaction•• 156

b. Study of the kinetics of the reaction 156 xi

CHAPTER PAGE

III. (CONTINUED)

2. !-Butyllithium 158

0 •. General VPC Analysis Technique 159

IV. SUMMARY 161

BIBLIOGRAPHY . 164

VITA •••• 171 LIST OF TABLES

TABLE PAGE

I. Preparations of Dialkylmagnesium Compounds from

Grignard Reagents . . • • • • . . . . • 31

II . Preparation of Dialkylmagnesium Compounds from

Commercial Magnesium Chloride ••••• 34

III . Preparation of Dialkylmagnesium Compounds from

Synthetic Magnesium Chloride 37

IV. Direct Preparation of Dialkylmagnesium Compounds from

· Alkyl Chlorides • • • • • • • • • • • • • • • 42

V. Solubility Data on Symmetrical Dialkylmagnesium

Compounds 47

VI . Preparation ?f Hydrocarbon-Soluble Organometallic

Complexes-Dialkyl�gnesium Compounds 52

VII. PMR Chemical Shifts in Selected Organometallic Reagents 53

VIII. Complexes of Dialkylmagnesium Compounds with

Triethylenediamine (TED) and MethyltriethyleneDiamine

(MTED) ...... 58

IX . Metallation of Aromatic Substrates by Dialkylmagnesium

Reagents ...... 61

x. Metallation of Heteroaromatic. Substrates by

Dialkylmagnesium Reagents ...... 64

XI. Metallation of Acyclic Substrates by Dialkylmagnesium

Reagents 75

xii xiii

TABLE PAGE

XII . Kine tic Data-Reac tion of �-Butyllithium wi th

Dimethyl Ether • • • • • • • • • • • • 104

XIII. Initial Homo logization Rates of 1M RLi with DME • 106 LIST OF FIGURES

FIGURE PAGE

1. Simplified Struc ture of Di -�-butylmagnesium Dimer ••• 46

2. E. S. R. Spectrum-Excess (�-Buty l) 2Mg Plus Benzophenone

in Hydrocarbon Solution (Violet Color) 84

3. Homologization of �-Butyllithium with Stoichiometric 2 Dimethyl Ether at 0°C.l/(Fraction Unreacted �-Butyllithium)

Versus Time (Minutes) • • • • • • • 90

4. Homologization of �-Butyllithium with Stoichiometric Dimethyl 2 Ether at 40 0 C; 1/(Fraction Unreacted �-Butyllithium)

• • • • • • • . . . Versus Time (Minutes) .• 91

5. Homologization of �-Butyllithium with Excess Dime thyl Ether

at 5°C; 1 /( Fraction Unreacted �-Butyllithium) Versus

Time (Minutes) ...... 92

6. Homologization of �-butyllithium wi th Excess Dimethyl Ether

at 40 0C; 1/(fraction unreacted �-butyllithium) Versus

Time (Minutes) 93

7. Homologization of �-Butyllithium with Excess Dime thyl Ether

Catalyzed by TMEDA (O.Ol34M) ; Fraction Unreacted

�-Butyllithium Versus Time (Minutes) • 97

8. Homologization of �-Butyllithium with Stoichiometric Dimethyl

Ether Catalyzed by TMEDA at 0°C; Fraction Unreac ted

�-Butyllithium Versus Time (Minutes) • • • • . 98

xiv XV

FIGURE PA GE

9. Homologization of !�Butyllithium wi th Excess Dimethyl

Ether at 0°C; Frac tion Unreacted !-Butyllithium

Versus Time (Minutes) • • • • • • . • . . • . • 101

10. ESR Spectrum-!-Butyllithium Plus Excess DME-77°K 102 CHAPTER I

INTRODUCTION

Al though dialkylmagnesium compounds have been known for over a . ' hundred years , relatively little effort has been directly expended on these reagents pe r !£; much of what we know about magnesium alkyls has been an outgrowth of previous studies concerned with the nature of the

Grignard reagent. Only recently has ·any effort been made to elaborate on methods. for preparing these reagents., to determine ·carefully their physical and chemical properties , and to develop.practical uses for these:reagents. It has been the purpose of the present study to review . the presently known methods of preparation of dialkylmagnesium compounds , to investigate some new preparative techniques for these reagents, to , examine their phys ical and chemical properties , and to carry out some basic chemical reactions involving their use . A summary of previous studies of magnesium alkyls follows.

A. METHODS OF 'PREPARATION OF DIALKYLMAGNESIUM COMPOUNDS

1 In 1929 Gi lman and Brown obtained a smal l amount of dimethylmag- nes ium as colorless crystals by a process which involved a di stillation of methylmagnes ium chloride at 190° under a vacuum of 0. 2 unn over a

2 3 period of nine hours. Gilman 's work confirmed von L�hrs and Flecks reports that the lower dialkylmagnesiums were solids and not liquids , as had been originally indicated in the mid-nineteenth century by Hal lwachs

1 2

and Schafarik4 and by Cahours. 5 The latter two workers prepared

dialkylmagnesium reagents of unknown pur ity from magnesium �etal and

the corresponding mercury dialkyls(l) , an example of the mo st venerable

of the indirect. synthesis techniques in organometallic chemistry .

R2H� + Mg -+ R2Mg + Hg (1) 6 Also in 1929, Schlenk and Schlenk discovered their now classi-

cal method of pre paring the magnesium alkyls, namely , precipitation of the dioxanate complex of the magnesium halide from an ethereal solu-

tion of Gr ignard reagent. From the varying amounts of ma gnesium halide

precipitated by this technique with different Gr ignard reagents , the

authors proposed equilibria (2) and (3) , wh ich they felt could exist

2RMgX � R2Mg + MgX2 (2)

.QI. �2Mg·MgX2 � R2Ms + MgX2 (3) in solutions of Gr ignard reagents. They proposed that ethylmagnesium

iodide consisted of 6EtMgi + 4MgEt2 + 4Mgl2 in solutio� and phenylmag-

nesium bromide of PhMgBr + O •.l lSMgPh2+ O.ll5MgBr2• Many authors since

then have been engaged in attempts to prove or disprove this hypothesis

7 8 9 10 and recent reviews, ' ' • of such wo.rk are available.

· The two indirect techniques, the Schlenks' dioxanate precipita7 11 12 13 14 15 16 17 18 19 tion ' ' ' ' and the older magnesium-mercury exchange ' ' '

have been and ar e still widely employed to pr epare dialkylmagnesium reagents in moderate to good yields. There are , however , several dis-

advantages in these techniques. Mercurial s are both poisonous and expensive. If the re action is effected without an ethereal type solvent ,

or in the ab sence of a solvating medium (that is , in the absence of a 3

Lewis base) , the magnesium metal-mercury dialkyl exchange method involves heating these reagents together in sealed tubes at rather high tempera ° ture s (115-120 ) for extended periods of time (24 hours or longer) . Such conditions preclude the synthesis of easily pyrolyzed dialkylmagnesium reagents. In addition, there arises the problem of separating the rea gents from by-product mercury, particularly if insoluble dialkylmagnes ium reagents are produced. The dioxane me thod is somewhat more rapid, and less stringent conditions are required, but even so, the specif ic pro cedure can not be called convenient . Further, the reagents so produced are solvated. They contain some residual halogen and are alway s diffi 2 cult to free of dioxane. 1

It would, of course, be most conv enient if magnesium alkyls could be prepared directly, particularly in a non-solvating medium, e.g. , much as solutions of lithium alkyls are prepared by direct react ion of the metal with alkyl chlorides in a hydrocarbon (4) . However, direct reaction

RX + 2Li -+ RLi + LiX (4) of magnesium with an organic halide in a ·solvating medium has always been considered to give the Grignard reagent , or perhaps as a better descrip tion, to give an organomagnesium reagent wh ich contains halogen . Until recently there have been relatively few reports of the direct reac tion in hydrocarbon of magnesium metal with alkyl chlorides , but alkylmagnes ium iodides were prepared in benzene as long ago as 1904 by Tschelinzeff.20

In 1931 Schlenk21 did investigate the preparation of a few different alkylmagnesium halides at room temperature in pur e benzene , i.e. , without the aid of solvating agents as "activators" or catalysts. Essentially 4

quantitative yields of £-butyl- and £-Octylmagnesium iodide , but only

a 55 per cent yield of £-butylmagnesium chloride were reported. 22 A group of Russian workers have repeated Schlenk's work and

reported that greatly improved yields (80-95 per cent) of alkylmagnes ium

halides could be ob tained from all the various normal al kyl iodides , bro-

mides , and chlorides from methy l to decyl in solvents such as isooctane

or dodecane by employing temperatures of 80-100° . By this direct route

secondary and tertiary alkylmagnesium halides could not be prepared,

presumably because of their thermal lability and the ease with wh ich

they undergo side reac tions at these temperatures . Bryce-Smith and 23 Cox have reported that Grignard reagents prepared by the direct reaction

in benzene contain cons iderably less than a stoichiometric amount of hal ide. It appeared to these workers that the products obtained vary in composition from a stoichiometry of R5Mg3I (or R3Mgzi) , for solub le pro-

duc ts derived from some £-alkyl iodides , to an almost halide-free hydro-

carbon insoluble product, derived from £-butyl chloride . The yield of

the £-butylmagnes ium reagent was indicated to be 80-85 per cent , based on the alkyl chloride , when an elevated reaction temperature was employed.

- This butylmagnesium reagent was characterized by slow precipitation of both active base and ha lide from the initial reac tion solution , the latter proceeding more rapidly than the former. For example, a freshly prepared

solution of £-butylmagnesium chloride in cumene , hav ing an initial active base to halide ratio of 2.0, slowly precipitated solid on standing .

After 19 and 50 hr . at 25°, the active base/ha lide ratios were 5.0 and 14 .6, respectively . The original concentration of magnes ium 5 produc t in solution decreased from 0.235 to 0.075 g. atom/liter during this time. The authors suggested the occurrence of a gradual change in composition with elimina tion of insolub le MgC 12 to give less ·soluble species until an almost insoluble R2Mg was ob tained.

A particularly interesting.n-amylmagnesium compound was reported by Glaze and Selman19 after the present study was initiate� . These workers reacted the-n-amyl halides directly with magnesium powder at approximately 85° over a period of several hours. Extraction of the reaction mixture, wh ich was of a "muddy" cons istency , with refluxing benzene for two hours left a solid "mud" phase and yielded the organa magnesium reagent in a very viscous , clear colorless liquid phase.

Removal of solvent benzene from the viscous liquid left a solid; addi tion of more benzene to the solid regenerated the viscous liquid, and the quantity of halogen in the liquid was not changed by th is process.

In one such liquid phase which contained a total of 0.17M amy lmagnesium reagent , the quantity of halogen (chloride) corresponded to only £!• two mole per cent amylmagnesium chloride. When the halogen was .bromide , the relative amount in the liquid was greater , and was still greater with iodide .

Thus there are several previous reports which suggest that in the absence of a solvating medium , direct reaction of magnesium with an organic halide (particularly with chloride) can produce a mixture of diorganomagnesium reagent and magnesium halide along wi th organa magnesium halide (5) .

3nRX + 3nMg -+ nR2Mg + nMgX 2 + (3-2n) RMgX (5) 6

Another "direct" route to unsolvated dialkylmagnesium reagents

involves the addition of MgHz (or its precursors , Mg and Hz) to ethyl 24,25 ene. Al though there is no halide by-product formed in this method ,

yiel ds of product are low and high pressures and temperatures are

<�r�quired. Other isolated examples of the preparation 'of solvated dialkyl

magnesium reagents have involved reaction of alkyl -l�.ate.e. with magnesium 26 and magnesium iodide , and metal -metal exchange between diphenylmagnes ium

2 and n-butyllithium . 7

Some particularly elaborate , but very interesting, techniques have

been developed to remove both halogen andcomplexed s�lvent from organomag-

nesium hal ides . In one process , de solvation of ethylmagnesium chloride by

dropwise addition of the Grignard solution in ether to boiling to luene ,

was fol lowed by admixture of .the prec·ip:i.tated EtMgCl with ·triethylaluminum

in hexane to give a halide-free solution. Distillation was then used

to remove hexane and finally (at reduced pressure) to remove triethyl

aluminum and leave diethylmagnesium. 28 There is a patented process of

electrolysis of alkal i metal tetraalkylaluminum solutions using a mag

nesium anode and separation of the products (MgRz·2AtRJ) as above, by 29 fractionation under reduced pressure. Al so, interaction of magnesium

° with ethyl chloride and NaBEt4 in diglyme at 115 for one hour , is

reported to yield a mixed reagent , MgEtz•BEt3 , wh ich can be separated 30 by fractionation.

One method of preparing organic derivatives of magnesium is almost

conspicuous by its absence. It has been known for many years that organic

derivatives of more electropositive metals will undergo metathetical 7

exchange with halides of metals lower in the electrochemical series,

. 31 ' 32 ' 33 . including exchanges involving· organomagnesium halides (6). For

(6)

example, Glaze and co-workers34 reacted-�·butylmagnesium iodide prepared

in benzene with beryllium chloride to produce di-�·butylberyllium in 7

per. cent yield. There are, however, only a few recorded instances of

attempts to use this metathesis to synthesize organomagnesium reagents

themselves; in each instance the use was highly speci fic. Nobis and

Robinson35 added 1,3-butadiene to a mixture of finely dispersed sodium

and anhydrous magnesium chloride in dimethyl ether to produce a "di-

Grignard" butadiene dimer adduct. In an unsuccessful attempt to obtain

the corresponding optically active dialkylmagnesium compound, Smith36

reacted an optically active !.�alkyllithiumcompound with inagnesium bromide.

Tzschach and Hackert37 prepared magnesium bis-di-cyclohexylarsenide from

. lithium dicyclohexylarsenide and magnesium bromide in tetrahydrofuran

(THF).

An initial goal of the present study was to develop general

techniques by which alkyls of the alkali metals could be caused to

exchange with halides of magnesium and by which unsolvated dialkylmag-

nesium compounds could be obtained, e.g., as illustrated below (7).

RMSX + RLi -+ RMgR + LiX (7)

During this study it became apparent that further advantage of the

Schlenk' s equili�rium (2) and of direct reaction (3) could be taken

·in synthesizing dtalkylmagnes.ium compounds. 8

B. PROPERTIES OF UNSOLVATED DIALKYLMAGNE SIUM COMPOUNDS

The smal ler normal dialkylmagnesium compounds (C1-C4) and one I . branched reagent (diisopropylmag�esium) have been found to be solids

and to be essentially insoluble in both aromatic and aliphatic hydro-

carbons . There is evidence that longer normal chain reagents might

dissolve in aromatic hydrocarbons , bu t solubility properties of.�-

and-S-dialkylmagnes ium compounds (other than diisopropyl) were not

determined prior to the present study. 12 Shown below are qu antitative data reported by Strohmeier on the

solubilities in benzene of some of the lower homologs. These previously

report�d so�ubilities must be considered .. upper limits because they were ., -

determined by simply titrating the total base from hydroly sis of the

supernatant benzene taken from excess solid magnesium reagent. Solu-

bilities found this way include dissolved compounds containing magnesium-

oxygen bonds (e .g., ROHgR) as well as dia1kyl reagents. Other data

Reagent Found in benzene

(C2H5) 2Mg O.Ol6M

. (�CJHJ}2Mg O.Ol2M

·' (iso-CjH7) 2Mg 0.024M

(!1,-C4H9) 2Ms 0.024M

obtained by Strohme ier indi�ate that these reagents must be even less

soluble (by at least an order of magnitude) in an aliphatic hydrocarbon,

hexane.

Magnesium alkyls are soluble in ethers and other basic solvents ,

and, as previously indicated, have not generally been obtained free of 9 solvation. It is, in fac t the difficulty in obtaining magnesium.alkyls free of such base which has precluded studies· of the solubilities of the.unsolvated reagents with branched chain .alkyls or·with higher homo logs· of the pr imary alkyl& . For example, Strohme i.er had prepared the 2 reagents he studied by the dioxanate precipitation technique.1 In or der to remove dioxane from his reagents it was necessary to heat them under high vacuum over a long period of time . (The isopropyl reagent was heated with a bath temperature of 11S0 for 15 hours under a "Klebevakuums" pro duced with a three-stage mercury diffusion pump; other reagents he studied were de solvated at even higher temperatures.) Strohmeierfound this desolva tion technique was unsuccessful with a branched reagent, (diisobutylmag-. ne sium) and with a higher homolog (di·n -hexylmagnesium); these reagents underwent thermal decomposition before all dioxane·had been removed. ·

(In the present study it has been found that di-S·butylmagnesium undergoes thermal decomposition when a de solvation is attempted at temperatures even well below 100°).

The factor of relative instability of reagents due to chain branch ing , a factor which is well recognized in thermal elimination reactions in general , is an obv ious reason for desolvation difficulties with the branched compounds. The reason why a higher homolog (but primary) should be diffi cult to desolvate is less obvious; they ar e stab le at elevated temperatures

(at least to 150° and perhaps higher) for many hours. :It is probable though that the longer chain reagents ar e viscous liquids, whereas the smaller reagents become amorphous solid powders upon desolvation. For physical reasons , removing solvent from a viscous liquid is more difficult than removing solvent from an amprphous solid. 10

That the longer chain r�agents might be solub le in benzene is

suggested by the solubilities of the "Grignard reagents" which were

prepared directly in that solvent by reactions of alkyl ha lides and which were found to contain less than a stoichiome tric quantity of • 2 halide in solution after reaction. Bry�e-Smith and Cox 3 as well as 19 Glaze and Selman expressed the belief that the structure of an alkylmagnesium bromide or iodide. in benzene solut-i-On is polymeric with an t unmixed bridge bonding of the type illustrated in (8). It is known that

(8)

X in· the, progression from alkyl iodides to bromides to ch lorides; the· organometallic in solution contained less and less halide . It is therefore natural to assume that in this progression, the structure in solution begins to assume a polymeric, totally alkyl-bridge-bonded type, illus

trated in (9) Indeed, from X-ray crystallographic data, Weiss38,39

R l ..)l / ' ,"' ...... / " ' Mg Mg /Mg (9) . , / �/""' 'a R R X

suggested that the dry residues obtained by removal of solvent from

RMgX ethereal solutions of (R•CH3• C2H5; X=Cl, Br) were ac tually mixtures of dialkylmagnesiums and magnesium halides. (Earlier Kirrmann, 40 Hamelin and Hayes had studied Grignard solu tions by thermogravimetric 11 techniques and expressed the belief that even in ethyl ether, the organa magnesium halides are not represented by a single compound, RMgX, nor by a simple mixture of MgRz, MgX2 and ether.) The dialkylmagnesium reagents studied by Weiss were reported to. be in the form of infinite polymeric chains; in the case of dimethylmagnesium, a complete analysis of crystal line material was reported. The unit cell is orthorhomb ic with an Mg-Mg distance of 2 .73�; the me tal atoms are connected by pairs of methyl bridges (Mg·C of 2.24X) wtta•� .aLmost tetrahedral arrangement of four me thyl groups around each metal atom, cf. (9).

The self associated structures of magnesium alky ls can be classed as typical electron deficient oligomers. (The concept of electron defic ient bonding has been extensively considered elsewhere and it is not within the scope of this survey to review that conc ept.)

All of these observations leave it di fficult to rationalize one of the observations of Glaze and Selman on the organomagnesium reagent they prepared from direct reaction of amy l chloride with magnesium . The products of this reaction were refluxed with benzene for about two hours ; an upper viscous benzene phase was obtained. With one set of analytical data it was indicated that one viscous benzene phase they obtained con tained the equivalent of 0.17M R2Mg and very little RMgX (O . OlM) , as previously indicated. It appeared then to Glaze and Selman that a true solution of diamy lmagnesium in benzene was produc ed. A freezing point depression. measurement of the benzene indicated that the contained diamyl magnesium reagent was dimeric . That is the observation wh ich is difficult to rationalize. From the evidence and arguments favoring polymeric self 1 2 assoc iation of other dialkylmagnesium reagents and of organomagnesium halides containing limited halide, one would also expect an . amyl'reagent to be extensively associated, rather than a simple dimer. Also, if this reagent is not polymeric, its ability to greatly increase the viscosity of benzene seems somewhat unusual .

Whatever the exact degree of association of magnesium alkyls in the solid state or in hydrocarbon solution, solvation by basic solvents such as ethers and amines depolymerizes the reagents to struc tures containing just one or two R2Mg units .

C. SOLVATION AND THE PROPERTIES OF DIALKYLMAGNESIUM COMPOUNDS

In general, just as with pr eparative studies, much mo're has been done in considering solvents with Grigrtard reagents than with diorganomagnesium compounds . Since there is every reason to believe that solvation of the two types of compounds could be analogous, several pertinent findings with the organomagnesium halides will be briefly stated in this section. 4 In work with ethylmagnesium reagents, Storfer and Becker 1 conc luded from colligative property studles, and from reaction kinetics with benzo- 42 nitrile, that diethylmagne sium (0. 5-lM) is dimeric in THF . (Ashby- had indicated earlier that ethylmagnesium hal ides are monomeric in THF and 43 triethylamine .) Vreugdenhil and Blomberg, on the other hand, conc luded that at low conc entrations (O . OlM) diethylmagne sium is mononteric in diethyl ether . If both conc lusions are correct, this behavior is somehwat anal 4 ogous to that found by Ashby and Smith4. .for: Grignard reagents in. ether . 13

Generally, the Gri gnard reagents were found to be monomeric at low concentrations (0.001-0.0 lM) and dimeric at higher concentrations (0.2M 4 and above) . Recent colligative property studie� by Ashby 6 indicate, however , that dialkylmagnesium compounds in diethyl ether and in THF are less highly associated than are Grignard reagents. Thus , it is difficult to reconc ile the work of others with that of Storfer and

Becker , and it seems most likely that dialkylmagnesium compounds are monomeric , rather than dimeric , in bas ic solvents. Smith and Becker 4 have reached the same conc lusion . 5

Scala and Becker showed that a considerable effect on reactivity 47 results from coordination of Grignard reagents by basic solvents.

Reaction of ethylmagnesium bromide with benzonitrile was found to be slower in more basic solvents such as THF and ether solutions containing various amines than in ethyl ether alone. This decelerating effec t on rate wa s especially evident when di glyme was the solvent , the authors ascribing its great effect on reaction rate to the especially stable chelate-coordination between one molecule of the magnes ium reagent and two oxy gen atoms in the same molecule of diglyme . The 48 authors cited Normant's bel ief that success in preparing alkeny l

Grignard reagents in THF is due to the greater stability of these reagents in this more basic solvent. From this belie£ the generaliza- tion was made that Gri gnard reagents are most stable in the more basic 49 solvents. Vink and coworkers found that the optical activity of

(+)(S) -1-ethoxy-2 -methylbutane in benzene was grea tly enhanced by coordination with ethylmagnesium bromide; THF was found to return 14 the optical activity of the system to its original value pr ior to addition of the Grignard, wh ereas di-£-butyl ether did not effect this change.

The se findings are eas ily rationalized with the postulate that the magnesium reagent is specifical ly coordinated by the ether with the lowest steric requirements in the system.

The only complete struc tural analyses ava ilable on crystal l ine 0 Gr ignard reagents are on solvated reagents. Stucky and Rundle5 found that both ethylmagnesium bromi de and phenylmagnesium bromide exist as ident ifiable monomeric units coordinated to two molecules of ether. The authors were unsuccessful in attempts to remove solvation from their

Grignard reagents and retain crystallinity; amorphous solids resulted. 2 ,3 ,39,40 Th ese ob servatLons,. 1 . a ong WLt h ot h ers c i te d ear lier , 8 8 ma k e it seem logical to define "Grignard reagents" as including solvation, and make it seem promising to further explore techniques of produc ing isolable dialkylma gnesium compounds by desolvation of Gr ignard reagents. 51 Coates and He slop ob tained monomeric complexes (in benzene) of dimethyl- and of diphenylmagnesium wi th 1,2-dimethoxyethane (monoglyme) and with N,N ,N',N'- tetramethylethylene diamine (TMEDA) . The (CH3 ) 2Mg : TMEDA complex could be sublimed intact at 65 -75°/0.01 mm. (m . p. 97 -98°C) , while the corresponding monoglyme complex did not me lt but did decompose by dissociation above 1 20°. These authors also found that comp lexes containing � mo lecules of a monoether such as THF could also be isolated. 52 Zakharkin isolated volatile monomeric TMEDA complexes of diethyl -, 53 di-g-propyl -, diisopropyl -, and di -£-butylmagnesium . Coates and Ridley ob tained a crys talline coordination complex of di isopropylmagnes ium wi th 15

!-butyl cyanide . The low solubility of the dioxanate complex of 4 dimethylmagnesium in diethyl ether has been known for some time . 5

It has also been known for some time that basic solvents promote some type of ionization of magnes ium alkyls. In a relative ly 55 recent report'on dimethy lmagnes ium in ether , Vreugdenhil and Blomberg

plotted "molar conductivity" .

electrical conduc tivity by the molarity in magnesium reagent) Y!· molar concentration and obtained a direct linear relationship up to a

one molar concentration. Ab ove this concentration the linear rela-

tionship no longer held and a maximum i� Amolar values was ob tained

at about 1.7 M. These authors found much lower Amolar values for 56 these solutions than had De ssy and Jones earlier, and they ascribed

this discrepancy to the presenc e of oxygen impur ities during the earlier work.

Strohme ier and coworkers57, 58 also measured the equivalent conductivity of dialkylmagnesium compounds in ether and in THF . The ir findings were interpreted in terms of the compounds' behaving like monovalent electrolytes, forming "triple ions ," as indicated by . equilibria

(10) and' (11) . With th is interpretation, dissociation constants K1 in

Kl + �·,, � MgR + r (10)

(U)

THF and ether were estimated to be in the range of 10-8 to 10-9 and -11 -13 . 10 to 10 , respectively . Values of K3 were said to be much - 2 larger , £!• 10 in THF . In other words , the extent of ionization1of 16 magnesium alkyls in basic solvents appears to be quite small . Inter- estingly enough , ma gnesium bromide, a "salt ," showed slightly lower conductivity in these solvents than did mag�esium alkyls. The equivalent- conductivities of the organomagnes ium compounds, measured in dioxane , ether , THF and triethylamine , increased with increasing dielec- tric cons tant of the solvent and seemed to.Strohmeier to depend on the solvent 's electron donor ab ility. It is noteworthy here that Ebel 5 and Schneider 9 were able to cause benzylmagnesium chloride to ionize, i.e. , give the intense red color of the benzyl anion, by addition of hexame thylphosphoramide to a colorless ethereal solution. 0 Quite recently , Psarras and Dessy6 reported on the polargraphic behavior of magnesium alkyls in 1,2-dimethoxyethane and elaborated on the nature of the electrolytic oxidative and reductive processes . It was their conc lus ion that R2Mg reagents undergo a two-electron oxi dation +2 at a mercury electrode to R2Hg and Mg • A one-elec tron reduction of the reagents occurred, but only when R was a stab le anion precursor such as benzyl or allyl , and appeared to form R: and RMg• , wh ich decomposed to

R· and magnesium metal . 1, 2 From NMR studies 6 6 of organomagnesium reagents , Fraenkel and coworkers concluded that chemical shifts are the same for both the

Grignard and the corresponding dialkyl compound. Temperature dependent spectra of reagents with special symmetry properties (2-me thylbutyl- and · 2-phenyl-3 ,3-dimethylbutylmagne sium) indicated that inversion, at the a-carbon atom (magnesium-bound carbon) oc curs fairly rapidly and, according to Fraenkel , by a mechanism involving electroph ilic displace-

2 ment of magnes ium (SE ) . Rates of inversion were said to be strongly 17 63 64 solvent dependent . Roberts and coworkers , ' from NMR. work with neohexylmetal compounds , have postulated the inversion mechanism to be unimolecular and of the SE1 type . 7 Seitz and Brown65 presented proton and t! NMR spectral evidence

for the existence of intermetallic comp lexes of methyllithium and dimethyl 7 magne sium in ether , Li2Mg(CH3)4 and Li3Mg (CH3) 5. Exchange of ti between methyllithium and the complexes appeared to occur at about the same rate

as methyl group exchange .

Salinger and Mosher66 studied the IR spectra of methyl, ethy l and

pheny l Grignard reagents and the corresponding dialkyl- and diarylmagnesium

compounds in ether and THF . It was their conclusion that carbon-magnesium -l bonding absorption occurred in wide bands between 500 and 535 cm (R•CH3 ,. -1 C2H5) and between 365 and 383 em (R.C6H5)• Whereas there was no difference in the proton NMR spectra of the Grignard reagents and their corresponding disubstituted magnesium reagents in THF, IR spectra showed no ticeable differences in these two reagent types. There was spectral

identity between Grignard reagents and synthetic mixtures of dialkyl- maane sium reagents and MgX2 in THF . Additions of excess MgX2 to these solutions altered the shape of the absorption band in the manner expected

for an increase in the concentration of RMgX and a decrease in the concentration of R2Mg , as would be expected from the Schlenk equilibrium (2).

(2)

Because of the real possibility that this equilibrium ca� b) . ·

rapidly established and ma4ttained t.t�_der ·many reaet.ton cottd.i-elons, it

would seem particularly important to consider the relative reactiyity of

dialkylmagnesium compounds when consideration is given to the mechanism

of Grignard reactions . 18

D. RELATIVE CHEMICAL REACTIVITY OF MAGNESIUM

ALKYLS AND GRIGNARD REAGENTS

67 Holm compared the reactions of ,!!-butyl Grignard reagents wi th those of di-,!!-butylmagnesium in ether and in THF with various substrates: acetone , methyl ac etate, azobenzene , me thyl trifluoroacetate , 1-hexyne , and !_-butyl crotonate. . Using ether he found that , except for 1-hexyne , there was in general a £!· 400 fold greater rate of reaction by di-n- butylmagnesium than by a-butylmagnesium bromi de toward all the substrates .

The bromide reacted four time s faster than the iodide , and the' chloride ,

20-30 times faster than the bromide . Using THF he found reactivity of both types of reagents toward all of the substrates except azobenzene

(wh ere the rate was accelerated) was cons iderably decreased relative to that in ether. In addition, the reactivities of the two reagent types (except for the iodide) were roughly equivalent in THF . It is suggested here that these results can be interpreted as fol lows . During react ions in ether , the Schlenk's equilibrium was not rapidly maintained;

RMgX is less reactive than R2Mg , so rates wi th solutions of the l.aotter were faster. During reactions in THF, the Schlenk equilibrium was rapidly ma intained, so whichever reagent was put into solution , rates were char - ac teristic of the R2Mg , wh ich now because of better solvation by THF was less reactive than it was in ether .

The "expected" addi tion of dialkylmagnesium reagents to ketones is complicated by "side" reduc tion and enolization of the ketone even more than is the ,same reaction with Gr ignard reagents . If the mol� ratio of dialkylmagnesium reagent to ketone is 1:2 , the reaction rate is much 1 9 more rapid during the first 50 ger cent of reac tion of alkyl groups; the rate is said to fol low second-order kinetics up to this hal f-way point and the rate to decl ine .rapidly -thereafter. 68 It has been proposed that the alkylmagnes ium alkoxide formed (12) during the first half of the

R2Mg + ke tone� RMgOR' (12) reaetton·is coft"stderably less reactive than the initial dialkylmagnes ium 17 reagent . When the molar ratio of R2Mg to ketone is increased to 1:1

(so that the alkylmagnes ium alkoxide formed did not have to react), the evidence is that the ratio of produc ts (from addition, reduc tion and enolization) is not far from that ob served with RMgx.68 In recent dev- 69 elopments , House and Respess have presented stereochemical evidence for the difference in reactivity of alkylmagnesium alkoxides in these systems , and several alkylmagnesium alkoxides have been synthesized by 70 71 workers in England . ' 7 2 House and T�aficante postulated that although MgBrz doe s not catalyze the addition of alkylmagnes ium reagents to carbonyl functions, it does suppress the tendency of ethylmagnesium alkoxides to give the side products resulting from enolfzation and reduction. They felt that a basic alkoxide mo lecule would be removed from the reaction medium by the presence of the Lewis ac id, MgBrz. Another interesting opinion they expressed concerned a trans ient yellow color observed during reactions of magnesium alkyls with ketones , which they speculated might represent a charge trans fer complex (formed _from the ke tone and the organomagnesium compound) and which might be a true intermediate. The authors could not maintain the transient yellow color long enough to measure spectral properties of the speci�s respons ible. In a similar view , Guthrie, Spencer 20 73 and Wr ight obtained an intense red color on mixing benzoin in ·benzene

with an excess of solid ether-free dime thylmagnesium • . No significant

amounts of side products were formed; work up gave a 92 per cent yield

of 1,2-diphenyl-1 ,2-propanediol (the expected addition pro4t-J.c�) .and only

8 per cent of enolization produc t. From the same reactants in basic

ethereal solvents , they ob tained less of the former and more of the

latter products .

While no one appears to have obtained specific spectral char

acteristics of intermediate complexes formed during reac tions of

dialkylmagnesium reagents with ketones , several workers have reported

on such complexes from Grignard reagents. For example, S. G. Smith74

obtained direct- spectroscopic evidence for the formation of complexes

between some highly hindered diaryl ketones and methylmagnesium bro mide in ether. A new band (in addition to. the ketone band) appears

in the ultraviolet spectrum of the mixture , but both bands disappear

rapidly at 25° . These results led Smith and su75 to postulate a mechanism (13) in which ketone reacts with monomeric Grignard compound

in a fast step to produc e a complex , wh ich then rearranges in a first

order process to produce the produc t. Ke sl� RMgX + ketone· 4 � complex • XMgOR ' (13)

The following reactions , involving an association complex, have

also been represented as a mechanism for additon of Grignard reagents

to ketones. This specific representation , (14) , (15) , and. (16) , is that

of Ashby and Smith, 44 but the general concepts, and many details, have

often been presented by others . 21

(14)

G =monomeric Grignard species , e.g. , R2Mg , RMgX

slow ketone + G ) complex (15)

fast complex + G � products (16)

The rate determining step (15) is thought by Ashby and Smi th to

be di splacement by the ketone of one of the strongly solvated ether mole-

cules attached to the magnesium atom. The authors de scribe the complex

so formed either as one involving polarization of the carbonyl group

(1 7a) or as one derived from coordinate covalent bond formation and

R' R

�:::d)···Mg' (17a) /- I R' X

having ionic character (17b) ; such a di stinction would seem to have �� {--0---Mg OEt (17b) /' + 2 / � I R' 9 X

little physical meanirig , to the present wr iter . They explain the

lower reactivity of RMgX , relative to that · of R2Mg , with ke tones as being due to the Grigna�? reagent 's gr eater Lewis acidity , which leads

to stronger bond formation with ether of solvation and thus slower forma-

tion of the association complex. An interes ting point to consider in

this explanation is that the actual differenc e in ac idity between RMgX

and R2Mg seems never to have been measured. 22

It would seem somewhat more logical to presently assume that R2Mg is more reactive than RMgX in a general way , inc luding their relative reactivity toward complexation by bases , i.e., until there is contrary evidence there is no need to assume that R2Mg is more ac idic than RMgX .

Sufficiently basic solvents might actually tend to stab ilize the R2Mg reagent (as already suggested, cf. page l�),wh ile similar solvents could possibly even enhance RMgX reactivity by fac ilitating maintenance of the

Schlenk equilibrium.

Th is latter possibility has been realized by Wotiz and coworkers , 76 who found that halogen- free diethylmagnesium reacts three time s faster in ether with 1-hexyne than does ethy lmagnesium bromide . Use of triethyl amine as a solvent increases the rate of reaction of EtMgBr with 1-hexyne , 77 but not the rate of Et2Mg . Again , the postulate being stressed here is that in some cases , reactions of Grignard reagents can ac tually be those of dialkylmagnesium reagents made available by the Schlenk equilibrium. Th is is not a new postulate . 7 House and Thompson 8 studied the effect of adding only the first alkyl group of Et2Mg on the ratio of conj ugate to normal additon to a,e unsaturated ketones . They found that the order of both reactions in R2Mg and ketone is the same and that the presence of MgBr2 slightly enhanced normal additon. (W ith long reaction times and a large excess of R2Mg , di-addition to the ketones resulted, presumab ly due to reaction of the enolate salt derived from conjugate additon with Et2Mg .) Conj uga te addition of magnesium alkyls also occurs with a,8-unsaturated esters. 23

An interesting study was made of conjugate addition (18) of dibutyl

magnesium to the hindered ester ,!-butyl crotonate . 7� . Use of butyl-

magnesium bromide instead of dibutylmagnesium lowered the yield of

R2Mg + CH3CH===CHQOBu + CH3bH--cH��OBu (18) lj 6M gR

,!-butyl-3-methyl-heptanoate obtained from reaction of the ester •.

Addition of MgBr2 to the Gr ignard reagent lowered the. yield still

fur ther . The authors postulat�d that the mechanism of conjugate addi-

tion to an a.B-unsaturated carbonyl compound should involve only the

dialkylmagnesium moiety from the Grignard reagent .

Contrary to the later results ob tained by others with ketones ,

Gilman and Brown80 reported in 1930 that nitriles react faster with

Grignard reagents than with their corresponding 4iorg anomagnesium com 81 pounds. MUch more recently (1963) Citron and Becker made a rate study

with various �- substituted benzonitriles and diethylmagnes ium in THF , � but did not compare rates with Gr ig nard reagents . They found that a

second order rate equation , rate proportional to [Et2Mg] [ nitrile1 ,

best fit their data. The linearity of a Hammett plot , with a positive

rho value (+ 1.57) , obtained from this study suggested that the same

mechanism, (19) and (20) , is operative in ali the examples studied.

The rate determining step suggested (20) involves a dipolar species

and is represented as involving transfer of R from R2Mg in a complex

to carbon of the nitrile gr oup . Storfer and Becker had already reported

(19) 24

(20)

thatf as was the case in reactions of R2Mg with ketones , only one -half of the av ailable alkyl groups are readily utilized in reactions of R2Mg 41 with nitriles. Schiff bases also add R2Mg via a process wh ich readily 2 utilizes only half of the available R groups . 8

Reac tion with epoxides probably provides the clear�st example of the dialkylmagnesium compound being the better reagent . Side produc ts obtained along with the expected addition produc ts (alcohols) on reac-

tion of Grignard reagents with substituted epoxides include halohy dr ins and their hydrolysis products . Only the expected addition products are obtained on reaction of dialkylmagnesium reagents with these epox� 15,83,84,85 86 ides . On the other hand , Resves and Fine did ob serve that both reagents reacted with 3,3,3-trichloro-1 , 2-epoxypropane to produce the product from halogen-metal interconversion , no expected addition produc t being formed.

Another specific example of the Grignard reagent being less useful was found in carbonation of some alkynyl reagents. Gi lman87 discovered that 1-hexynylmagnesium halides gave low yields of acidic produc t on carbonation (<5 per cent) whereas dihexynylmagnesium (from Et2Mg and hexyne-1) gave a 73 per cent yield of 2-heptynoic acid. While it would be a mistake to generalize that the magnesium alkyls are superior for carbonation, the following is another example wh ich can be so interpeted.

Apparently, after findin&: poo-r results from conventional carbonation of'iJ:. 25 88 an ethereal solution of (+) (S)-2-methylbutylmagnesium chloride , Lardicci

obtained a 71 per cent yield of (+ ) (S ) ac id produc t on carbonation of the

suspension obtained by slow addition of the Grignard reagent in ether to

boiling xylene followed by continuous distillation until the distillate

was ether-free . There can be little doub t that the indicated removal of

ether converted the Gr ignard reagent to a mixture of magnesium bromide

and dialkylmagnesium.

A particularly interesting reaction was studied by Coleman and 89 Blomquist. These workers obtained 90-100 per cent yields of n-butyl-

amine and no ammo nia on react ion of (Bu) 2Mg with ClNH2 • When BuMgX was used, less amine was formed, ammonia was ob tained, and the yield of amine

decreased according to X in the order Cl>Br.>I. Formation of amine (21)

apparently involves. a coupling process (22) followed by me tallation. If

R2Mg + NH2Cl -+ RNHMgC l + RH (21)

RMgX + NH2Cl -+ RC l + XMgNH2 (22) such coupling is restricted to the dialkyl reagent , then only to the extent

that the Schltmk ' s equ.ilibrium produces R2Mg could ·RMgX produce amine (23) .

Mg + NH Cl RMgCl + + RNHMgCl Rz 2 + RNHz -+ RH (23) Coupling of Grignard reagents with organic halides has not been widely employed, and coupling of the dialkyls, even less . Zakharkin and

coworkers90 ,91 showed that although interaction between organomagnesium

compounds and alkyl halides in ethers of high solvating power results mainly in the formation of typical Wurtz reaction products, suggesting

a heterolytic reaction process, the same reactants in non-solvating 26 media, cumene , resulted in the formation of large amounts of bicumene ,

suggesting a free-radical process .

E. METALJ:.,ATION AND POLYMERIZATION BY MAGNESIUM ALKYLS

In the past most of the chemistry of the simple organometallic

reagents , e.g. , lithium alkyls and Grignard reagents (i.e., not

1T-complexed me tals), has been classified as "carbanion"· chemistry .

Such reagents, R-metal, have been looked upon as salts of very weak - ac ids , R-H, and as sources of very strong bases, R: , carbanions .

Re lative kinetic and the�odynamic stabilities of these simp le organo metallic& have bee� thought to be a function of relative carbanion stability. Foremost amongst the reactions wh ich characterize these organometallic& as "carbanionic" has been the hydrogen-metal inter

change with various substrates , wh ich reaction has been variously

called "metallation," "ac tive hydrogen reaction," "solvolysis ,"

"Zerewitinof reaction," etc. The process can be represented with a general substrate , Z-H, as follows (24) .

R-metal + Z-H � R-H + Z-metal (24) Probably the most useful of these exchanges is that of aromatic hydrogen with lithium alkyls , c ommonly called simply" Metallation :·(25) .

R-Li +· Ar-H � R-H + Ar-Li {25)

This type of metallation, wh ich has been widely employed for syntheses of lithium aryls, has never been ob served wi th magnesium reagents , presumab ly because the latter are not sufficiently "carbanionic" to . react with an "ac id" as weak as Ar-H . 27

Diorganomagnesium compounds are known to react readily with many 51 3 such active-hydrogen containing compounds as amines , arsines , 6 alco- 71 92 hols , and silanes . The latter reaction with silanes leads to the

formation of RMgH . The cleavage of ethers by organometallic reagents 93 also probably involves metallation. Podall and Foster reported that

diethylmagnesium underwent cleavage react ions wi th diethyl ether , diglyme ,

and other ethers much more readily than does MgH2 . On the other hand , it 92 appeared to Bauer that RMgH cleaves diethyl ether more rapidly than does

R2Mg . Products from ether cleavage by magnesium alkyls have not been fully studied, and the reaction has apparently served no useful purpose.

(A useful cleavage of dimethyl ether by lithium alkyls was found in the

present study.)

Alkynes are probably the weakest ac ids for which magnesium alkyls have served as useful metallating agents. Podder and coworkers94 expressed

the reac tivities of R2Mg with 1-hexyne in terms of rate constants corre- ponding to a two-step, competitive , consecutive , second order mechanism with the first half of the alkyl groups again being more reactive than the second half. Reactivities of alky l groups were found to be in the order isopropyl> ethyl>g-propyi >methyl. Since this order correlated with the number of B-hydrogens on the alkyl groups , "anionic hyperconjugation" 94 was suggested to increase the reac tivity of R2Mg (26) !

(26) 28

In a more conv entionalattempt to rationalize the relation of struc 9 ture to reactivity , Dessy and coworker s 5 suggested that a four-center ,

' non-ionic transition state 6ould be postulated for dialkylmercury-dialkyl- . , magnesium exchange reactions (27) in ethereal solvents . Salinger and 96 Dessy applied a study of equilibrium positions in this metatheti cal

reaction to the determination of carbanion stabilities .· They reported

an or der of decreasing stability to be phenylethynyl , phenyl , me thyl ,

ethyl , isopropyl, and said that this order would be expected from con-

sideration of inductive (+I) effects.

(27)

Polymerizations initiated by organometallic reagents are also

generally considered to involve carbanion chem_istry . Apparently the

first use of a magnesium alkyl for polymerization of ethylene came in

1958; Podall and Foster reported the followi ng observations . 93 Diethyl-

magnesium in hexane or heptane reacts quantitatively wi th ethylene at 100°

� and £!• 700 p.s.i. to produce (after quenching wi th water) liquid hydro-

carbons ranging from C10 up and a small quantity _of solid polyme� expected

products fr om additio.n polymerization. Ether retards addition (the chain

growth processes) wi th diethylmagnesium, whi le the ethy l Grignar d reagent

does not add at all to ethylene or other simple olefins. Diethylmagnesium

was also found to react ( 27) wi th bis -(diphenylene) ..eth ylene .in ether

dioxane at 25°C to give after reaction wi th benzyl chloride , 1-ethyl-2-

benzyl-bis- (diphenylene) -ethane •. Again, Grignards failed to undergo this r aqdition.

(27) 29 97 More recently Langer stated in a patent that he had polymerized ethylene to solid polyethylene with a catalyst prepared by reacting equimolar quantities of di-�-butylmagnesium and N,N,N' ,N'-tetramethylethylenediamine . He also used heptane as diluent and a temperature of 100°C, and a pressure of �· 700 p.s.i. After a reaction time of three hours , · a yield of twenty grams of polymer per gram of di-�-butylmagnesium was reported by Langer . He also reported that EtMgCl, Et2Zn, and Et3Al did not give solid polyethylene under the same conditions . Most 98a recently (1966) Fontanille and Sigwalt polymerized styrene , d.•methy l- styrene , and 2-vinylpyridine , in the presence of diphenylmagnesium in hexamethylphosphoramide . Yields of polymers were quantitative . The elec tronic ab sorption spectrum of the reaction mixtures during polymerization indicated the reaction took place through polystyrene carbanions ("living polymer"). CHAPTER II

RESULTS AND DISCUSSION

A. PREPARATION OF DIALKYLMAGNESIUM REAGENTS

The preparative techniques developed in the present study involve (1) exchange reac tions of alkyllithium reagents with alkyl magnesium chlorides and with magnesium chloride , and (2) direct reac tion of magnesium me tal with alkyl chlorides in a hydrocarbon or in a hydrocarbon-ether solution.

1. Preparation of R2Mg by Me tal-Metal Exchange between RLi and RMgCl

� MgCl2

Alkyllithium reagents were found ·to react readily with alkylmag nesium chlorides or with magnesium chloride in ethereal media to yield the corresponding dialkylmagnesium reagents in solution and precipitated

RL i + RMgCl R2Mg + LiCl (2-1)

2RLi + MgC l2 R2Mg + 2LiCl (2- 2) lithium chloride . Table I shows the compounds produced (1) from Grignard reagents when the main solvent was diethyl or di-n-butyl ether; yields were essentially quantitative . Diethylether wa s removed from the initial reac tion solutions by continuous codistillation with hydrocarbon solvents , i.e., hydrocarbon solutions were prepared by "desolvation" of ethereal solutions .

Ether- free solutions of magnesium alkyls were also prepared by exchange (2) of magnesium chloride with lithium alkyls directly in hydro carbons . With comme rcially available anhydrous magnesium chloride , the

30 TABLE I

PREPARATIONS OF DIALKYLMAGNESIUM COMPOUNDS FROM GRIGNAim REAGENTS

Reactants Product1 R2M2 Orig. !Uf&Cl RLi cone . Compound Milli- Milli- ·in .. After Desolvat ion Prepared R Solvent moles R Solvent moles ether Solvent Cone .

b (_!!-Bu) 2Mg .n-Bu Bu20 84 _!!-Bu heptane 84 0.84M benzene . l.OM

(iso-Bu)MeMg iso-Bu Et20 81 Me Et20 82 0. 76M benzene o.7 2M 2 (�-Bu) 2Mg �-Bu Et20 90 �-Bu C6H1 2 89 0.78M cyclohexane 0.9 M a (!-Bu) 2Mg !,-Bu DME 100 1!,-Bu hexane 98 0.46M hexane- 0.68� cyclohexane d (iso-Bu)2Mg iso-Bu Et2o 78 iso-Bu Et20 77 0.42M cyclohexane O.lM e 2 f (!_-Bu) tfg -t-Bu Et20 48 t-Bu · pentane 79 0.8 M benzene

8Solvent was 5 per cent dimethyl ether in cyclohexane . b solution contained di-n-Bu20·

CFrom this solution there was obtained a 1.35M (�-Bu) 2Mg solution in benzene by vacuum distilling the hexane cyclohexane and redissolving the residual viscous liquid in benzene . ·w ..... TABLE I (CONTINUED)

' Owhite precipitate formed; N.E. 66, calculated for (1!2-Bu)zMg , 69 .

eThis amount of !-butyllithium added to reac t with !-BUMgCl and with sglid · by-product MgC1 2 present . fsolution contained ethyl ether. Heating to 80°C caused thermal decomposition with the formation of a black precipitate .

w N 33 exchange is slow and inc omplete . Addition of ether to the - reaction mixture apparently "activates" the magnes ium chloride and definitely fac il itates the exchange reaction. After exchange , the ether wh ich had been added could be removed by distillation of the hydrocarbon solutions . For the exchange to proceed to completion, addition of at least two equiva lents of diethyl ether per mo le of alkyllithium and attritional stirring of the hydrocarbon reac tion mixture wa s required. Tables II and III show the results of some alkyllithium-

MgC l2 exchange runs.

If the ether employed for activation is dimethy l ether and the reaction is carried out at room temperature without high .speed stirring , a side reaction can occur . Methy lene insertion into secondary or tertiary alkyllithium reagents occurs , forming a homologous primary reagent

(2�3). These new primary alkyllithium reagents then react wi th MgC l2 in

3 2CH3CH2·q8-Li + CH30CH3 -� CH3CH2CH2CH3 + Li0CH3 + CH3CH2·CH·CH2Li ( 2- ) 6a3 tu3 the usual manner (2-4). The homologiza tion process is discussed further

(2-4) in part G . following .

Seyeral processes were tried for activating commercial magne sium chloride so that it would exchange with lithium reagents in the absence of an ether . Some success was realized in treating the salt with an ether or an alcohol or some other "_ac tivat ing" agent , followed by a complete desolvation. (These processes are described in detail in the experimental section , Chapter III.) Some activating agents are rather tightly he ld, such as the "active ethers ," e.g. , dime thyl ether (�) •

TABLE II

PREPARATION OF DIALKYLMAGNESIUM COMPOUNDS FROM COMMERCIAL MAGNESIUM CHLORIDE

s Reactants Final d H&Ctz RLi Reaction Reaction Per. Cent Cone . Compound Amount AmQunt Time Temp . � Yield R2Mg o . Prepared (moles) ·solvent R (moles) Solvent (hr.) ( c) R2 (m/1) . ' Mg

· · a b c d - e (!,-BU) 2Mg 0.22 . none .. !,-Bu 0.36 n-hexane 16 10 77 l.68 g (!,-Bu)2Mg 0.498 none !,-Bu 0.80 ,!1-hexane f 27 25-35 65 2 .33 h (!,-Bu)2Mg 1.168 none s Bu 1.94 ,!1-hexane 22 0-25 5s-8 8 k (!!,-Bu)2Mg 0.27 none !!,-Bu 0.11 benzeneJ 2 0-10 90 0.2 1 s e (Neopentyl)2Mg 0.1378 none Neopentyl 0.07 !!,- pentane 2 0-10 79 0.27 m n .( s -B\1)2 Mg 0.15 cyclo- !,-Bu 0.11 cyclohexane 16 25 2l 0.05 hexane ° (!,-Bu)2Mg 0.5 benzene ·!,-BU 0.026 cyclohexane 16 25 61 0. 13 p (!,-Bu)2Mg 0.07 toluene s-Bu 0.11 ,!1-hexane 4 30 76 0.15 q r (!,-Bu)2Mg 0.2 tetralin s-Bu 0.12 !!,-hexane 20 25 77 0.22

___p (2-MeBu) 2Mg toluene 2-MeBu 0.07 benzene 2 40 95 0.13

w � 8Anhydrous powdered MgCl2 . TABLE II (CONTINUED}

b . DME , 0.44 mole, added.

cHigh speed stirring employed.

dunless otherwise indicated, yields were based on analyses of aliquots of supernatant reaction solutions for total base, active alkyl and magnesium .

eEther of complexation present in quantities equivalent to R2Mg .

fDME, 0.46 mole, added at start; then an additional 0.45 mole DME added after q hours.

SRepresents a mixture of 2-butyl and 2-methylbutyl products due to some homologization. h DME, 2.1 moles, added.

iYield based on ratio of �-butane to 2-methylbutane obtained on hydrolysis · (VPC} .

J100 ml . ether added.

konly 50 per cent of product soluble.

lPrepared by homologization of £-butyllithium with DME. m Anhydrous lump MgC12 ball-milled in 100 ml . cyclohexane . nPer cent of total alkalinity represented by magnesium containing compounds .

0Anhydrous powdered MgCl2 stirred vigorously overnight with ethyl ether; then desolvated with beJ}zene.

w PBall-milled lumps treated with isopropyl alcohol; then desolvated with toluene. V'l TABLE II (CONTINUED)

�Cl2 "6H20 dehydrated with isoamyl alcohol ; then desolvated with tetralin. rper cent of total alkalinity represented by active alkyl content.

8Concentration after distillative desolvation.

w 0\ TABLE III

PREPARATION OF DIALKYLMAGNESIUM COMPOUNDS FROM SYNTHETIC MAGNESIUM CHLORIDE

k Reactants Final b MaClz !lti Reaction Reaction Per Cent Cone . Compound Amount Amount Time Temp . Yield R2Mg O Prepared (moles) Solvent R (moles) Solvent (hr.) ( C) RzMg (m/1)

a c (!_-Bu)2 Mg 0.25 benzene s-Bu 0.40 cyclohexane 2 25;..35 90 0.5 d (!_-Bu)2Mg 0.05 .benzene s-Bu 0.26 cyclohexane 0.25 25-35 95 0.55 e c (!,-Bu)2Mg o.o8 benzene t-Bu 0.12 !!.-pentane 2 25 95 0.5 f (!!,-Bu) 2Mg 0.09 benzene !_-Bu 0.026 cyclohexanf7 0.25 25-35 958 0.5 h (!_-Bu)2Mg 0.14 benzene !_-Bu O.l8 benzene 0.75 25-35 70 0.42 h (.!l-Bu) (!_-Bu)Mg 0.14 benzene !_-Bu 0.07 benzene 0.75 25-40 80 0.5 n-Bu 0.07 i j (!_-Bu) 2Mg 0.6 cyclohexane !_-Bu 0.89 n-hexane 5 40 83 0.4 - awashed solid by-product from direct reaction of !_-BuCl and Mg in benzene-DME .

bunless otherwise indicated, yields were based on analyses of aliquots of supernatant solutions for total base, active alkyl and magnesium .

c w Ether of complexation present in quantities equivalent to R2Mg . -...J TABLE III (CONTINUED)

- dwashed solid by-product from direct reaction of �-AmC l and Mg in benzene . ewashed solid by-product from direct reaction of iso-PrCl and Mg in benzene-DME. fwashed solids from distillation under vacuum of ether from �-BUMgCl in ether .

8Based on negative Gilman Color Test IIa at end of reac tion period. hne solvated solid produc t from reaction of �-BUMgCl with benzyl chloride .

ine solvated solid product from reaction of �-BuMgC l with Cl2. jBased on per cent of total alkalinity represented by magnesium containing compounds . kconcentration after distillative desolvation.

w CXI 39 or tetrahydrofuran, and require higher distillation temperatures for removal than ethyl ether requires, For example , distillation with ' benzene will not effect the removal of DME from. the crystal structure of MgCl2 , whereas it rapidly removes ethy l ether.

No activating process produced a magnesium chloride which is as active or which reacts as completely with alkyllithium reagents

in ether-free hydrocarbons as that magnesium chloride obtained as a � solid by-product from reaction or preparation of an organomagnesium reagent , · e.g., from reaction of a Grignard reagent with benzyl ·Ghloticle ..

It woul� appear that a salt with a particular crystal structure is the most important factor in reac tivity toward the exchange (2). X-ray analysis- of these "active" MgC12 residues from reactions would no doubt supply some interesting ·information . iri this reg�d.

Reaction of an alkyllithium reagent in a hydrocarbon with an

''ac-tiveM.'�sium chloride is the most important of the techniques used to produce unsolvated dialkylmagnesium reagents . This rapid,. low tem perature exchange technique has permi tted the preparation of hydrocarbon soluble, bu t relatively thermall�· unstable reagents such as di-�- and di-�-butylmagnesium in excellent yields . Thus , one may react directly with an , alkyllithium reagent , the MgC l2 produced as a by-product of a direct magnesium alkyl· preparation, or that produced by reaction of a

Grignard reagent wi th a reactive chlorine-containing compound (see

Table III)· , or less desirably , that produced by ether-activation of commercially available anhydrous MgC l2 (see Table II) without removal of the complexed ether in the MgC l2 . If an unsolvated dialkylmagnesium 40

reagent is desired, either the MgC l 2 mu st be further treated to desol-

vate it by continuous di stillation wi th a hydrocarbon solvent prior to

exchange With alkyllithium, or the magnesium alkyl mu st be similarly

desolvated after exchange . Desolvation of the magnes ium chloride is

preferable because it is not as sensitive to pyrolysis as are the

dialkylmagnes ium reagents . Ac tually , reac tion of uns olvated ac ti-

vated MgCl 2 with !-butyllithium is the only route to uns olvated di-!-

butylmagnesium, since this latter reagent decomposes readily dur ing

the desolvating distillation procedure .

If one employs the Mg Cl2 produced (2-5) as a by- product of the

2RC 1 + 2Mg -+ R2Mg + MgCl2 � (2-5)

direct preparation of a hydrocarbon-soluble primary dialkylmagnesium

reagent (see following paragraph) , there is no ether present in the

process to begin with , and desolvation techniques may be dispensed

with entirely . Use of MgCl 2 produced in such a manner (2-5) for the

exchange reaction ( 2-2) in effect allows preparation of two different

unsolvated dialkylmagnes ium reagents utilizing the same initial mag-

nesium source. There is considerable versatility in this exchange

procedure; it would appear to be amenable to the synthesis of a variety

of dialkylmagnes ium reagents.

2. Direct Preparation of R2Mg �Magnesium Metal and Alkyl Chlorides

A variety of techniques were employed to remove magnesium. chlo-

ride from dialky lmagnesium reagents prepared by direct reaction (2-5)• of

magnes ium with alkyl halides . Some of the results are summar ized in

' I ' 41

Table "IV� One technique developed was the use, as solvent for this direct reaction , a hydrocarbon containing a limited amount of an ether .

When diethyl ether wa s used under those circumstances , magnes ium chloride stayed in solution. However , its precipitation could be forced by removal of ether through codistillation with a hydrocarbon solvent , or by vacuum stripping to dryness and redissolving the R2Mg in hydro carbon. If, instead of diethy l ether , dimethyl ether was emp loyed in the hydrocarbon solvent for reaction of magnesium with an alkyl chloride, MgCl2 did precipitate almost completely (>90 per cent) du ring the reaction (2-5) . The remaining soluble portion of the MgCl2 could be precipitated du ring a distillation desolvation step. It thus appears that ethers other than dioxane may cause reversible formation (and precipitation) of MgCl2 from RMgCl · solutions .

AnotbeI r technique was developed to separate MgCl2 and R2Mg after direct .preparations in hydrocarbon solvents wi th no ether present. ' For example , preparati�n of di-�-butylmagnesium in benze�e fram magnesium and �-butylchloride was readily carried out . Both products of the reaction are insoluble. The organometallic produc t, however , was dissolved preferentially by passing gaseous dimethyl ether into the slurried react ion mixture. Desolvatio� was then effected by the usual method of continuous codistillation with benzene unt il the residual mixture was free of DME .

It has been found that yields of R2Mg are generally improved when ethers are emp loyed ,as cosolvents in these direc t preparations , especially when magnesium in the form of turnings is used. Elevated TABLE IV

DIRECT PREPARATION OF DIALKYLMAGNESIUM COMPOUNDS FROM ALKYL CHLORIDES

. Reactants RCl, Mg, Reaction Reaction Final Per · Compound Amount Amount Solvents Time Temp. Cone . Final Cent i . oc Prepared . mi:>le- g. atom Hydrocarbon Ether Hr. R2Mg Solvent Yield

a (.!_-Bu) 2Mg 0.1 0.1 Cyclohexane DME 2 Days 25 0.39M Cyclohexane- 79 DMEd

. (,!.-Bu) 2Mg 0.4 0.4 none Et20 4 35 0.85M Benz.elle- 79 ethere b (,!.�Bu)2Mg 0.39 0.415 Benzene Et2o 3 77 0.3M Benzene- 79 ether£ c (.!!2,-Pr) 2Mg 6.4 0.4 Benzene DME 3 35-45 0. 7M Benzene- 80 DME

a 100 inl . DME used as first solvent for reactio.u. This DME gradually displaced by 100 ml . � cyclohexane , but not completely . N TABLE IV (CONTINUED)

b Amount of et�er, 0.48 mole.

CAmo�nt of DME, 0.�4 mole.

dsolution contained approximately 1-2 molar equivalents of DME per mole of R2Mg .

eBenzeile added to resid*e (after- evaporation under -vac�um) containing :�gCli and . R2Mg . Final mo lar Ether:R2Mg ratio = 0.7.

£Molar ratio of ether to RMg c�pounds was 2:1 after di stillation of 50 per cent of solvent; molar ratio of RMgCl :R2Mg was 4:1.

8DME bubbled into a slurry of (�-Bu) 2Mg and MgC 12 in cyclohexane until solution complete.

hTwo volumes of 50 ml . each of benzene added with stirring and refluxing for 1 and 2 hr ., respectively. 1 Unless otherwise indicated-, yields were based on analyses of aliquots of supernatant solutions for total base, ac tive alkyl and magnesium .

.p. w 44 reactton temperatures also shorten reaction time and , give. impreved · yields of RZ� · Commercially av aila�le m�nesium powder .is not reac tive at . all, but .undoubtedly it could so111ehow be a:ctivated·prior to react�on with alkyl halide. It has .been . suggested that the .magnesium be freshly powdered prior to use . and Mendel99 describes such a prep aration from regular Grignard turnings. As Bryce-Smith and Zakharkin observed , it was found in the present study that secondary .and ter tiary , alkyl halides did not·react (2-5) with magnesium in hydrocarbon solvents alone and required the pres ence of ethereal cosolvents to form the desired products.

In one attemp t at another technique for the direct . preparation , one which is actually a modification of the .exchange� reaction , �butyl chloride was · caus�d to react (2-6) with a mixture of lithi.um and magnes ium metal in ether to af ford .a 60 per cent yield of di-�-butylmagnesium. ·

2BuC l + Mg + 2Li � Bu2Mg + 2LiCl+ (2-6)

B� PHYSICAL PROPERTIES OF DIALKYLMAGNESIUM REAGENTS

As . indicated in Chapter I, knowledge about the physical properties of dialkylmagnesium reagents is . limited . Previously me�sured solu bilities of dialkylmagnesium compounds in. hydrecarbons , were . listed on page 31; in all of those cases , the dialkylmagnesiums investigated were infusible solids which possessed extremely low solubiliti�s in. hydrocarbon solvents. It is known that these compounds can be poly� merically associated, and this factor may influence tQe nature of their physical state.as well as their solubility characteristics in some .cases. · 45

For example, it is known that �-butyllithium and �-butyllithium are liquids at room temperature and are completely miscible in hydro carbon solvents . Al though the analogous di��-butylmagnes ium is an infus ible solid at room temperature , and reportedly possesses a low order of solubility in hydrocarbon solvents, the present study has shown that di -�-butylmagnesium is a liquid and is , like· �-butyllith ium, miscible in all proportions with hydrocarbon solvents. Di-�-butyl magnesium is also extremely soluble in hydrocarbon media.

Utilizing the C-Mg and Mg -Mg bond distances reported by Weiss38 and us ing tetrahedral configurations around each Mg atom , it was noted that although it is possible to cons truct mode ls of polymeric di iso propylmagnesium , di-�-butylmagnesium,, or diisobutylmagnesium, it is not possible to cons truct associated forms of di -�- or di -!-butyl magnesium beyond the dimer state (cf. Fig. 1) . Th is restricted ability to self-associate is apparently due to the steric hindrance imposed by the presence of a- and a-methyl groups in the former compound and three a-methyl groups in th� latter. Dimeric magnesium alkyls would correspond in mo lecular size to the tetrameric oligome rs previously found for �-butyl lithium and !-butyllithium. Presumab ly the ability of magnesium to form tetracoordinate structures permits the ready formation of polymers as soon as steric hindrance is relieved, and th is condition pr.evai1.s· in the �-butyl and isobutyl structures.

Table · ·v shows some solubilities of symmetrical dialkylmagnesium compounds observed in the present study. Although Glaze has found that di.,.�-'amy lmagnesium is somewhat soluble in benzene (and we corroborated Me Me '---� J c, &' ' ;/ \--M� H ' Me ' ...< ! cu · 'H � , \ /;M,___ c \ / .4 '\ . -H c ? u_V 'H a_V...... ?' r Me

Figure 1. Simplified structure of di-�-butylmagnesium dimer.

ol:'- 0\ TABLE V

SOLUBILITY DATA ON SYMMETRICAL DIALKYLMAGNESIUM COMPOUNDS

Solvents Ob se�ed Concentrations '.• R:zMg Hydrocarbon Ether . moles/liter CODDDents .

a (,!.-Bu) Mg Benzene none 0.43 Compound is a liquid 2 a n-Hexane none 0.39 and exhibits infinite solubility in these # solvents

(i so-Bu)2Mg Cyclcihexane none 0.09

(i so-Pr) 2Mg Benzene none 0.1 Compound is a solid. a (2-Mebutyl)2Mg Toluene none · 3.8 a (Neopentyl)2Mg n-Pentane DME 0.28

((CH2)4Mg)x None Ether 0,07 Analogous Li compound is very soluble ()2M) � ...... 8Not necessarily upper limit of solubility•

�etastable system. 48 \ this finding) , in our hands , the solution as first obtained was quite viscous , but it thinned out on standing for a period of from 16 to 24 hours . Since Glaze reported that di ...u-am ylmagnesium was dimeric , it seems probable that the solid product when first dissolved in benzene is of a higher associative order , but that these oligomers gradually

"break ·up11 and "dissolve" to form dimers . There is somewhat more steric hindrance to the formation of polymeric chains in this compound

than in di �n-- butylmagnesium. Probably of greater importance, however, is the lower van der Waals ' energy between polymeric chains in di-n- amylmagnesium due to the progressively more random dis tribution of side chains in the polymer as the carbon number increases .

A metastable solubility of di-n-butylmagnesium was found during its ether desolvation with cyclohexane . Whe n this reagent was prepared in the ab senc e of ether, it was found to be a solid of low solubility in hydrocarbon. Ether does affect solubilization of this reagent in a hydrocarbon, just as it does other insoluble magnesium alkyls studied.

However , as a di·n-butylmagnes ium cyclohexane solution containing ether was distilled, and the ether con.tent.· was reduced to only traces, the reagent stayed in solution. Eventually the solid reagent did come out

. of solution, very slowly , but at one stage the cyclohexane was over

0.2 M in di�n-butylmagnesium.

It now seems safe to assume that with sufficient solvation , magnes ium alkyls are monomeric , and that desolvation (2-7) in a

desolvation• 2(R2Mg)ether � R4Mg 2 + ether+ . (2-7) hydrocarbon can produce a soluble dimer . The process of polymeric

association (2-8) produces an insoluble reagent. From the above

nR4Mg 2 :t (R4M8 2) n (2-8) observations on diamyl and di-�-butylmagnesium, it appear s that

the polymerization process (2-8) of magnes ium alkyls is not neces

sarily a fast one, and that the structure of the alkyl groups deter mines:rates as well as positions of equilibrium in th is process.

Although bis(2-methylbutyl)magnes ium was previously reported to be insoluble in hydrocarbon solvents , 62 in our hands it was found to be extremely soluble in toluene and to form a quite fluid solution.

Th is compound, as well as di-neopentylmagnesium , probably prefers to exist as a dimer , again due to ster ic hindrance to further associa tion •. Perhaps not surpr isingly, tetr·amethylenemagnesium . has a low order of solub il ity even in ethereal media , although its lithium counterpart is quite soluble .

It was found that hydrocartion-insoluble dialkylmagnesium com pounds could be readily solub ilized by solutions of organolithium reagents in hydrocarbon solvents. This mu st result from formation of interme tallic complexes. The complexes did appear to have a pre ferred stoichiometry . For example, �- and �-butyllithium dissolved enough di-�-butylmagnesium (in the presence of excess undissolved magnesium reagent) to form 1: 2 complexes , i.e. , to have the composi tion (�-Bu) 5Mg2Li and (�-Bu)4 (�-Bu)Mg2Li in solution . Disruption of the tetracoordina te polymeric structure of the magnesium reagent by the alkyllithium reagent takes pla�e (2-9) leading to lower molecular

(2-9) 50

weight, soluble oligomers • . Presumab ly electron-defic ient three-center

bonds take part in the intermetallic structures;

Another interesting discovery was that mixing a soluble dialkyl

magnes ium with an insoluble dialkylmagnes ium resulted in a hydrocarbon

soluble complex containing both types of al kyl groups . The stoichio

metry in this case is not as clear-cut as in the intermetallic com-

·plexes . Ratios of alkyl groups in the complexes were some times found

close to 1:1, but often were not found in ratios of whole numbers . The

· mixed dialkylmagnes ium compounds may be formed by simply mixing· a hydro

carbon solution of the soluble R2Mg with the insoluble solid one (e.g. ,

. !-butyl-isobutyl reagents) or, they may be formed (2-10) by react ing a

RMgX + R'Li -+ RMgR ' + LiX + (2 ..10)

Grignard reagent having one type of alkyl group with an organol ithium

reagent possessing the other group. Since the reaction is usually

carried out in an ethereal medium , desolvation with a hydrocarbon sol

vent mu st be carried out .

It was di scovered, rather surprisingly, that reaction of isobutyl

magnesium chloride with methyllithium fo llowed by filtration to remove

lithium chloride and desolvation of the filtrate with benzene to remove

ether resulted in a somewhat viscous solution of the mixed me thy liso

butylmagnesi�. It is known that ne ither of the two individual magnesium

alkyls is soluble in benzene. On cooling , a clear glassy second layer

separates; warming redissolves the glass . An explanat�on for this

behavior can once again be found , not so much in steric hindrance in

the mixed alkyl to the format ion of polymers, for there is none apparent 51

in methylisobutylmagnesium , but rather in the ability of the polymeric chains to allign themselves to form high molecular weight crystalline

solids. This "loose" structure permits solvent attack to break down

the chains into smaller , more soluble oligomeric units . The methyl

isobutylmagnesium solution maintains its viscosity with time (unlike di�n-amylmagnesium) ; the reagent is obviously in a higher state of association than dimeric.

Table VI shows some data on mixed magnesium alkyls. Another

indication of the formation of mixed organomagnesium compounds was obtained when, upon metallation of resorcinol dimethyl ether in benzene cyclohexane solution, an insoluble crystalline compound formed (2-11)

(2-11) containing equivalent quantities of 2 ,6-dimethoxyphenyl and �-butyl groups. Metallation is discussed in part D following.

Listed in Table VII are NMR. chemical shifts in the "organo me tallic region" for a number of reagents examined in the present

study . The shifts , given in cps relative to benzene , lie in the region

six to eigh t ppm upfield from benzene , where signals from protons on carbon geminal to metal or in methyl groups vic inal to metal generally are found. Chemical shifts of mixed magnesium reagents , as well as

those of symmetrical reagents , and of lithium alkyls are shown from both hydrocarbon and basic media. The signals from basic media always appear upfield from those in hydrocarbons , and the shift is much greater with magnesium alkyls than with lithium alkyls. This particularly effec

tive shielding by base of the magnesium reagents is entirely consistent TAB� VI

PREPARATION OF HYDROCARBON-SOLUBLE ORGANOMETALLIC COMPLEXES DIALKYLMAGNESIUM COMPOUNDS

Reactants ComJ!l ex . · Approximate: Final Mobi't:' .Rat io .. Complex R2Mg Solvent R'Li or ll'.�M& Solvent :Cone . R:.lt'. Group� · " Prepared (Amou'4t) ·-: . (Amount) (Amount)· (Amount) m/1 in complex te: �-Bu) 5Mg2Li (,!l-Bu) Benzene n-BuLi ,!l-Hexane 0.77N (3 t1llllO e�) (5 ml.) {1 mmole) cyclohexane (0.825 ml.)

1 �-Bu)4(.!.-Bu)Mg2Li (!i�Bu)teg Benzene ,!_-BuLi eycl ohexane 0.81N 4.0

(3 mmo es) 1 (5 ml .) (l mmole) (0.825 ml .)

(�-Bu) 2Mg- (�-Bu) 2Mg none (.!,-Bu)2Mg a-Hexane- 0. 78M 1 (,!.-Bu) 2Mg (2 .5 umoles) . (3.4 umoles) cyclohexane

(iao·-�u) 2Mg- (iso-Bu) 2Mg none (�.-Bu) 2Mg Benzene 1.3M 1 (�.-Bu)2Mg (5.4 mmoles) 8 .: · " " , .• · 2 2 - �-Bu) 2Mg• (!,-�u) 2Mg .28-BW.i +. : : Benzene !_-BuLi + Benzene 0.5 M 0.8 MgCl2 - (19 ml .) MgCl2 - (90 ml .) 35 mmoles (35 mmoles- (,!l-Bu) 2Mg (!_-Bu)2Mg) b (iso-Bu)2Mg · (Me) 2Mg �-Bu-MgCl Ether MeLi Ether o.72 M 0.8 0.18 moles 0.18 mo les

VI �Some loss of (�Bu) 2Mg·probably occurred due to heat of reaction of MgCl 2 and !_-BuLi . N bsolution in benzene . Reaction mixture in ether desolvated by continuous codistill4tion with benzene. 53

TABLE VII

PMR CHEMICAL SHIFTS IN SELECTED ORGANOMETALLIC REAGENTS

Chemical Shifts (cps from· Benzene) Reagent Solvent CH3 .. RCH2 R2CH'. (CH3} 3C

. {B.-Bu) 2Mg Cyc lohexane 406 b (!l· BuhMg THF/TED 474 c d (neoPent) 2Mg Pentane/DME 447 360 e (2-MeBu) 2Mg Benzene 425

(!-Bu} 2Mg Cyc lohexane 436 b (,!-BU} 2Mg THF/TED 455 f ·

(.!!2-Bu) (Me)Mg Benzene 470.5 425 b .(iso-Bu} (Me}Mg THF/MrED 543 469

.(iso-Bu) (!_-Bu}Mg Benzene 447.5 355.5 b (iso-Bu) (S-Bu)Mg THF/MrED 469 384

(iso-Bu}3MgLi Benzene · 450

{t-Bu)4MgLi2 Benzene 366.5

.. (!-Bu) (£-Bu)Mg Benzene 442 . -� �-- 356

.(!-Bu) (B.-Bu}Mg THF/TEDb 470 377

.(B.-Bu)3 MgLi Cyc lohexane 442 f (,!-Bu}MgCl Benzene/Et20 426 -- -- 54

TABLE VII (CONTINUED)

Chemical Shift (cps from Benzene) � Reagent Solvent CH3 RCH2 R2CH (CH3)3C

(l'-Picely1)2Mg I Cyclohexane · 327 (2,6-Di�thoxyphenyl)THF 457 (.!,-Bu)Mgg

(Phenyl) (S-Bu)Mg THF 381

MeLi THF 553

MeLi THF /MTED 556

. n-BuLi Cyclohexane 482 . n-BuLi THF 493 - n-BuLi THF /MTEDb 497 - .!!2,-BuLi Benzene 480

.!!g,-BuLi THF 491 d neoPenLi Benzene 483 362 .. .!,-BuLi Benzene 493

. iao-PrLi Benzene 485 - f .!!2,-PrLi Benzene/Et2o 490 t-BuLi Benzene 372 .5 -

. asymbols : THF, tetrahydrofuran; TED , triethylenediamine ; MTED, methyltriethylenediamine ; DME , dimethylether.

bTED/or MTED/organome tallic reagent ratio of 'one .

cDME/R2Mg · ratio ef four •

. ds-Bu group separated from metal by me thylene group . 55

TABLE VII (CONTINUED)

eBroad unsymmetrical multiplet. fEt20/organometallic reagent ratio of two.

8Proton para to Mg, 22 cps downfield from benzene . 56 with the process of solvation, the reverse of the process represented in equation ( 2-7) , being one which dissociates oligomeric reagents.

The presence of the Lewis acid MgC1 2 leads to a downfie ld shift in the proton resonance of a group adj acent to magnes ium. Even in the presence of ether (a Lewis base) , the me thine proton in �-butyl magnesium chloride (or di-�-butylmagnesium-MgC12) dissolved in benzene resonates 10 cps lower than it does in di -�-butylmagnesium-benzene solutions alone .

The signal s from the mixed compounds corroborate the conclusion, based on solubility studies , that the magnes ium alkyls form true inter metallic compounds , i.e. , not simply mixed solut ions in wh ich each reagent retains its structural inteirity. It can be noted that in the n-butyllithium- di-n- butylmagnes ium system, (n-Bu) 3MgLi, only one u;le .thy lene triplet appeared, ' centere.cl-446 cps from benzene . Also , only one singlet appeared for the !-butyl protons' in the !-butyllithium di-!-butylmagnes ium sys�em, (!·Bu)4MgLi2 . These results indicate either that there is a rapid exchange of all alkyl groups in the com plex in solution, or that there is only one environment for each alkyl group in the mixed compound. In either case � there must be a new environment provided by a mixed compound, because the one signal it provides is not at the average position provided by the separate reagents . For examp le , if the signal from (n-Bu) JMgLi, found at 442 cps, were the average from

Similar evidence for compound formation is seen in comparing the two chemical shifts from the unsymmetrical dialkylmagnes.ium rea gents with those from the corresponding symmetrical reagents. For example, there is an upfield shift of 36 cps for the protons on the methylene group adj acent to Mg on replac ing half of the n-butyl groups in di-n-butylmagnesium with !,-·butyl groups . These shifts suggest an increased shielding of the me thylene protons due to inductive effects in the !,-butyl group. As might be expected, there is a concommitant decrease of 10 cps in the chemical shift of the !,-buty l group from its normal position in di -!,-butylmagnesium .

C. COMPLEXES OF DIALKYLMAGNES IUM REAGENTS

WITH BRIDGEHEAD DIAMINES

Table Vl!t shows some of the results obtained from reactions of triethylenediamine (TED) and me thyltriethylenediamine (MTED) wi th various symmetrical and mixed dialkylmagnesium compounds. Both reagents could cause precipitation of complexes . The triethylenediamine complexes were only slightly soluble in hydrocarbon solvents while the methyltriethylene diamine complexe s were moderately soluble and could be easily recrystal lized. The TED complexes all me lted above 200° (with decomposition) while the MTED complexes melted considerably lower. All of the comp lexes tested, however , decomposed on me lting . For example , the TED derivative of di-�-butylmagnes ium melted at 201- 203° (dec.) , wh ile the MTEP: deriv ative melted at 160°C (dec.).·. With :TED and MTED, di-�-butyl- and di-!, butylmagnesium formed only insoluble 1: 1 complexes , wh ile with TED, TABLE VIII

COMPLEXES OF DIALKYLMAGNESIUM COMPOUNDS WITH TRIETHYLENEDIAMINE (TED) AND METHYLTRIETHYLENEDIAMINE (MTED)

&211 Audne COUlDlex ·u Aaeuftt , Ameunt ,Selubility NB - m.p. Batie* Reagent mm.oles Solvent Reagent mmoles in Me dium Foun� The�y (dec.) Amine:R2Mg

w c ° (,!-Bu)2Mg 2 .13' Cyclohexane TED 2.07 Lo 248 250 2 n-2o3 1.0 2.33 Cyclohexane MTED 2.38 High ------160° 1.0

{!.-Bu) 2Mg 3.0 Benzene TED 3.0 Low ------1.0 2.5 Benzene MTED 2.5 High 309 264 163-165° 1.0 -- , 2.2 Cyclohexane TED 2.1 Low ------1.0 . (!!,-BU) 2Mg 1.3 Cyclohexane TED 0.65 Low ------0.5 1.0 Cyclohexane · MTED 1.0 No ,l;nso.lflble �omplex Formed ·

(!!,-Am) 2Mg 1.0 Benzene TED 0.5 No Insoluble Complex Formed

(iso-Bu) (Me)Mg 2.8 Benzene MTED 2.9 High ------115-270° o.5b c (iso-Bu).(!.-Bu)Mg 1.52 Benzene MTED 1.52 High ------140-165° l.O

8Ratios determined by analysis of NMR spectrum of complex dissolved in THF .

bRatio Me to iso-Bu groups = 3: 2.

CRatio · iso-Bu to !-Bu groups 3.2:2 V1 = 00 59

di�n-butylmagnesium formed insoluble 1:1 and 1: 2 complexes , am ine :rea-

gent mole ratio. Such results are taken to indicate that sufficient ·

amine dissaciates oligemeric reagents and precipitat�s the manomer, as .

illustrated in equation (2-12) , but that with a limited amount , the

intact dimeric reagent is complexed, as illustrated in equation (�-13) .

R4Mg2 + 2TED -+ 2R2Mg:TED (2-12) 1:1 complex

Ri.Ms2 + TED -+ Ri.Mg2 : TED (2-13) 1:2 complex

No insoluble complex was formed betwe�h di.,.n-butylmagnesium

and MTED or between di-a-amylmagnesium and TED.

The mixed alkyls of methyl-isobutylmagnesium and isobutyl-!-

butylmagnesium reacted with MTED to form non-stoichiometric materials

with wide melting point ranges ; slow decomposition occurred through-

out this range . The MTED-methyl-isobutylmagnesium product contained

a high ratio of alkyl groups to MTED (average of 4 alkyl groups to 1

MTED) while the isobutyl -!- butylmagnesium product contained a ratio of

about one alkyl group to one MTED. However, in both cases· there was

less of the isobutyl group than either methyl or S-butyl groups . It

is probable that both the isobutylmagnesium and the me thylmagnesium

moieties act like di-a-butylmagnesium with TED in that a 1: 2 amine :R2Mg

complex may be formed. It would also appear that , like the di-n-butyl-

magnesium-MTID case, the di-isobutylmagnesium •MTED complex is more solu-

ble than either the dimethylmagnesium- or di -!- butylmagnesium•MTED

complexes . The results ob tained support the conclusion that a mixed

dialkylmagnes ium reagent does not necessarily form a mixed dialkyl-

magnesium- amine cemplex, but instead the · ainine ·may disproportionate the 60

reagent · and form symmetrical dialkylmagnesium-amine complexes . Rapid

interchange of alkyl during comp lexation by the amine must allow this

.. result, which· :I.e apparently caus�ci·:by precipitation _9f the''.least· .soluble

.. .• c;;pmplex. as it iS -formed during :;th±s. :rapid inter.C,hail�e_· of alkyl'groups

It ·certainly ·.could be expected· that the· least sal�ble complexes would .

indeed be the most symmetricaL ones.

D. METALLATIONS WITH DIALKYLMAGNESIUM REAGENTS

Metallations of several different substrates were attempted with

dialkylmagnesium reagents . Tables IX:,. :and X show the general conditions

employed and the results obtained with some aromatic substrates. It is ·

apparent th at the magnesium compounds do not metallate aromatics with

the facility that lithium reagents do (as was known) , but the results

obtained unequivocally demonstrate that the metallation by magnesium

alkyls does occur . The se results, illustrated in the following equations,

would seem to be the first reported metallations of aromatic and hetero-

aromatic substrates by magnesium reagents .

R2Mg (2-14) 0. O MgR s cS�s R

R Mg 0 2 (2-15) t)MgR CH30 OCH3 �J OCHJ R2� . � - (2-16. ) ' '() � TABLE IX

METALLATION OF AROMATIC SUBSTRATES BY DIALKYLMAGNESIUM REAGENTS

Reaction Reactiol) Apparent Metallation Yield Reagent Source8 Solvent Substrate Catalyst Time(hr.) Temp . (°C) Product (%)

- b c <.!.- �u) 2:tofg I Toluene Toluene TMEDA 2 112 Benzyl�

- f c d (!_-Bu) 2Ms I Benzene Benzene TMEDA 4 80 Benzy lMgR. 14

<.!.� Bu) 2:tofg II Hexane- RDE� None 16 75 2 ,6-Dimethoxyphenyl- 50 cyclohexane MgR h . \ i j (!:-Bu)2 :tofg II Hexane- RDE8 THF 16 25 llone 0 cyclohexane 16 75

(!:-Bu) Mg II Hexane- RDE8 THF + 1 25 6-Dimetho:qphenyl- 57 2 k Z,._ . • cyclohexane K- (�-Bu) K n g c j

(!:-Bu)2 :t-Jg II Hexane- RDE8 TMEDA + 16 75 2 ,6-Dimethoxyphenyl- >90 cyclohexane Li- (�t-Bu) l Li-n,m

- 8I Reaction of �-butyllithium with commercial anhydrous MgCl2 in he�ane-dimethyl ether 0\ solution , fol�owed by partial or complete desolvation. ,_.

- TABLE IX (CONTINUED)

!!-Reaction of �-butyllithium with synthetic "active" de solvated MgCl 2 (from (�- BUMgCl + Cl2)) in hexane-cyclohexane . bPartially de solvated di-�-butylmagnesium solution in hexane further desolvated by codistillation with toluene . Some pyrolysis wi th formation of white precipitate was noted.

cTwo moles TMEDA per mo le of di-�-butylmagnesium . d produc t analyzed by NMR assay of dried ethereal · solution of mixture of acids ob tained on carbonation .

eHexane and cyclohexane distilled under vacuum , toluene added. No pyrolysis observed.

£Hexane distilled and replaced by benzene . Some pyrolysis noted.

gResorcinol Dimethyl Ether.

hNMR analysis showed no organomagnesium compounds in solution and the insoluble crystalline product (_d issolved in THF} to consist of an equivalent amount of 2 ,6-dimethoxy phenyl- and �-butyl-groups linked to magnesium.

1Two moles THF per mole di-�-butylmagnesium present du ring first 1 2 hours; six moles THF per mole present during next 16 hours .

jNMR analysis showed no loss of di-�-butylmagnesium and no appearance of me tallated resorcinol dimethyl ether.

kRun number 6 to which was added 2 moles potassium !-butoxide per mo le di-�-butylmagnesium.

lRun number 8 to ·Which was ·added :at least ·. 2 moles lithium !.; butoxide :per mole di-s butylmagnesi�.

0\ � TABLE IX (CONTINUED)

lDNMR analysis showed ab sence of (.§_-Bu) zMg in product.

nProduct analyzed by NMR assay of DzO solution of mixture of sal ts of acids obtained on carbonation (NazC03) added to aid solubilization) .

0'1 w TABLE }t METALLA.TION OF HE�ROARO!fAT IC SUBSTRATES BY DIALKYLMAGNESIUM REAGENTS

� Reaction Reactio� Apparent Time Temp . ; Metallation Yield Reagent Source8. Solvent Substrate Catalyst (hr. ) (°C) Product (%) ·' .. ,:· ' ' � ,. . (!.-Bu) 2M& II Hexane- . y-Picoline None 2-3 25 y-Picolyl.Mg-Rb ...75 cyclohexane

(!.-Bu) 2Mg II · Hexane... Thiophene None · 20 25 None� 0 cyclohexane 33 c:i.-Bu) 2Mg II Hexane- Thiophene None . 20 75 2-ThienylMgRd - cyclohexane: e (!.- Bu) 2M& II Hexane- Thiophene THF 1) 12 75 Nonec 0 cyclohexane 2) 16 25 e (!.-Bu) r"& II HexlU1e- . Thiophene TBF <1 25 2-Thienyl-Kg -90 cyclohexane K- (Ot-Bu) f h i (!.-Bu) 2Mg III . Cyclohexane Thiophene TMEDA Few. Min. 25 None 0

(!.-Bu) 2Mg IV TBF Thipphene . TED 24 70 "None.c 0 TED · complex c (!-Bu)2Mg v TBF Thiophene TED 24 70 .None 0 TED complex

"' �- B�)4

.-r.. ·· TABLE X (CONTINUED) .

a i - Reaction of �- butyllithium with commerc�al anhydrous MgCl2 in hexane- dimethyl ether solution followed by partial or complete desolvation. '

II - Reaction of �'"' butyllithium with synthetic "active" M&C l2 (from (!_- BuMgCl + Cl2)) in hexane.,Cyclohexane.;

' III - Reaction of magnesium turnings and �- butyl chloride in cyclohexane-ether , followed by desolvation .

IV - Reaction of �- butyllithium with synthetic "active" MgC12 (from �- butylMgCl + ben�yl chloride) in cyclohexane , followed by percipitatiQn with triethylenedi�ne.

V - Reaction of !.- butyllithium with synthetic "active" MgCl2 (from �..,. butylMgCl + benzyl chloride) in cyclohexane , fo�lowed by precipitation with triethylenediamine • .

VI - Solution formed by add ition of �- butyllithium in C6H12 to a slurry of �- Bu) 2Mg in benzene . b � analysis indicated a.3:1 ratio .of metallated to unmetallat�d y-p�coline . Product identified by derivatization with benzoyl . chloride to y-pbenacylpfridine.

cNMR analysis showed no loss of di-�-butylmagnesium and - absence of me tallated thiophene.

dNMR analysis showed no (!- Bu)2Mg and presence of metallated thiophene in supernatant solution ; NMR analysis of D20 solution of . salts of acids obta�ned on carbonation of the reaction mixture showed a 2 :1 iatio of the salts _ of 2-methylbutyric and 2-thienoic acids . e �o moles THF per mole di-�-butylmagnesium present during first 12 hours ; six moles THF per mole present during next . l6 hours . £ Previous run with THF to wh ich wa s added 2 moles of potassium �- butoxi� e per mole of Cl' di-�-butylmagnesium. VI

. .. TABLE X (CONTINUED)

B product analyzed by NMR assay of the D20 solution of the mixture of salts ob tained on cafbonation _(Na2C03 added to aid solubilization) . h. 'Two moles THEDA per mole of R2Mg. � showed absence of - 2-thienoic acid on workup of acids derived from carbonation of product.

� showed presence of �- butyl- -and 2-thienyl- , .but no �-butyl-organometallics .

0\ 0\ 67 Unlike metallations carried out by alkyllithium reagents , which ·

are catalyzed by Lewis bases such as THF and tetr�thylethylenediamine

(TMEDA) , me tallatians by dialkylmagnedum reagents are actually sup,--

pressed by· these bases. This observation is particularly clea-r.· in

results obtained with thiophene and with ·resorc±nol dimethyl ether.

Both are metallated by di·�-butylmagnesium in refluxing hydrocarbon

solutions ; neither is metallaLed if excess THF or TMEDA is present in

the solutions .

These results are not consistent with interpretation of magnesium

reagent reactivity in the simple terms of carbanionic character . Surely

polar basic solvents should facilitate release of a carbanion from R2Mg,

so the fact that such solvents do not facilitate metallation tends to

mitigate against any s·imple nuc leophilic proton abstrac tion process as

a mechanistic pathway for metallation with these reagents. It does

seem poss�ble that metallation could involve dimeric reagent and

nuc leophilic displacement on one magnesium of the bridging alkyl group

. by the species to be me tallated • . As illustrated below, this process

could proceed through formation of a n-complex (2-17) , rearrangement

to an ion-pair (2-18) , and elimination of hydrocarbon (2-19) .

___,:..(17)

�

(R = �-butyl) (18) (19) � � slow fast 68 Such a process is made more plausible by the known -rapid inter-

change of alkyl groups as well as by the high degree of nuc leophilicity

of the magnesium reagent . This process would be inhibited.by electron-

donating reagents such as THF and THEDA, wh ich complex readily wi th the monomeric magnesium reagent and thereby stabilize tt, i.e., . remove

electron-deficient dimer from the reaction medium. In short , the elec-

tron-deficient ·· character or 1T-acidity of dimeric magnes ium alkyls· seems

to provide a better basis of interpreting reac tivity than does carbanionic character.

While the mechanism speculated on above is illustrated, steps (17),

(18 ), and (1 9), as involving the ma-.ntenance of paired electrons , this. is done for simplicity , and because illustration of any further details would be even more speculative . However , the possibility should be considered that charge transfer does occur during this reaction , possibly in forming the initial complex, step (17 ). If this is so, then single electron transfer (SET) processes do not require basic solvents ; 106 as has been previously _ suggested.

The results shown in Table IX suggest that the second alky l group is not normally displaced from the metallated product initially formed.

This is probably due , not only to steric hindrance, which is obvious in the case of the me tallation product of resorcinol dimethyl ether , but also to stabiliziation by 1T-electron donation from the aromatic ring to metal in the produc t.. In the case of resorcinol dimethyl ether (DMR) , additional internal stabili�iit�en of the me tallated product is afforded � .. .. 69 by the presence of adjacent methoxy groups . Th is added stabilization prevents pyrolytic decomposition of the spec ies ; th is produc t remained unchanged even during extended reflux at 70-80° , used to attempt. utilization of the second alky l group in me tallation.

�I�� -R

The same stabilizing effect was found to be true for solutions of the dialkylmagnesium reagents themselves in the presence of added

Lewis bases such as THF, TED, and TMEDA . Thus even after extended refluxing at 70-80° , di-,t-butyl- and di-�-butylmagnesium did not eliminate olefin in the presence of these bases. However, in the absence of bases (or of readily meta llatable substrates) the !_-butyl reagent decomposed readily even at room temperature , wh ile solutions of di-�-butylmagnesium pyrolyzed slowly at reflux temperatures of

70-80° . Stabilization of R2Mg by solvating base is almost surely some function of the quantity of base. In the desolvation of hy drocarbon solutions by distillation, decomposition was not observed until the concentration of base present was ·.reduced to- the same molar level as the R2Mg present. Then , with further di stillation, it was apparent that elimination set in , presumably according to the following equation.

RMgH + Olefin (2-20)

Di-�-butylmagnesium which has partially undergone this thermal elimination (pyrolysis) possesses the ability to effect some me tallation of toluene , even in the presence of Lewis bases . If pyrolysis I

of the dialkylmagnesium reagent had not previously been effected,

me ta llation of toluene could not be realized in the presence of bases .

Benzene was also partially me tallated wh en desolva ted (part�aUy· ···

pyrolyzed) di-�-butylma gnesium in benzene . containing two equivalents

of TMEDA was heated to reflux for 4 hours. -7hese results are parti-

cularly interesting because benzene and toluene are relatively. unreac-

tive substrates .

A possible explanation for these results , which show an enhanced

reactivity Gf: pyrolyzed R2Mg , lies in formation ( 2- 20) of a soluble alkylmagnesium hydride wh ich· can retain some self-associated form even

in the presence of basic solvents. It is · a fact that me tal hydrides

and alkylme tal hy dr ides are often more strongly associated and less

affected by base than are me tal alkyls . A good example of this fact

is the trimeric existence of R2AlH wh en R3Al is monomeric , e.g., wh en

!-·Another exa'inp le ·is the retention :of B2H6 in its 108 dimeric form even in basic solvents . Some me tal hydrides , e.g. , LiH ·

and MgH2 , are not even sufficiently solvated by base to be solub le 109 . therein . Tentatively it is postu lated tha t structures such as ( 2 -21 ) and ( 2- 22) can be formed by pyrolysis of magnesium alkyls. It is

further postulated that these structures represent hydrocarbon soluble

H /\ ' R---Mg Mg� I I H H

\jg F. (2- 21) (2-22) 71 reagents whic h may retain their self-associated form, at least in some finite equilibrium concentration, even in the presence of base.

These postulates , and related phenomena previously discussed, are reiterated with the illustrative equations 'shown below, wh ere B represents an ether or tertiary amine . The equations show some prob- ab le states of association and solvation and are not intended to represent all possible (or even probable) states .

R4Mg2 + 2B -+2R2Mg :B monomeric a-complex (2-23) 6 ) R4Mg 2 minus R3HMg2 soluble hydride (2-24) olefin

R3HMgz + B-;R3HMg2 :B dimeric 1T-complex (2-25)

dime ric + unsaturated __,. 1T-complexed (2-26) reagent substrate reaction intermediRte

While the postulates concerning oligomeric alkylmagnesium hydrides are quite speculative if based only on the present experimental results , they do have rather good analogies, and they are amenab le to future experimental tests. It seems most probable that solution processes which completely di ssociate elec tron deficient oligomers e.g. , a-comp lexation of monomers as in (2-23) , decrease the reactivity of me tal alkyls and metal hydrides . It is because of their electron deficient struc tures that species like R4Mg 2 or R3HMg2 can initiate metallation through 1T -coordination of ArH, as in (2 -26). 72

Solvation of these self-associated species intac t,. As in (2-25) might

even promote their reaction, as does solvation of self-associated

lithium alkyls.

Observations made with hydrogenolysis reac tions provide inter-

esting evidence both that lithium alky l solvation increases reactivity

(n-complexation) and that magnesium alkyl solvation removes reactivity

(a-complexation) . Hydrogenolysis of lithium alkyls , which can be

looked upon as a metallation of hydrogen (2- 27) , requires base. In

the present study , magnesium alkyls were found to resist hydrogenolysis

RLi + H-H -+ RH + LiH (2-27) with or without base . In an earlier study Screttas obtained rapid

and complete hydrogenolysis of �-butyllithium in the presence of very

sma ll catalytic quantities of a Lewis base, such as 0.01 mo lar equiv lOOb alent of TMEDA. Surptisingly, it was found in the present study that

_!-butyl lithium wou ld not undergo hydrogenolysis in the presence of a 0.5 molar equivalent of TMEDA and 1.0 mo lar equivalent of di-,!-butylmagnesium .

Since �-butyllithium undergoes very rapid hydrogenolysis (2-27)

in the presence of this much TMEDA , it appears that the di-�-butylmagnesium

. coordinates more strongly wi th TMEDA than does ,!-butyllithium . The coordination which occurs in the case of the di alkylmagnes ium reagent is different from that of the alkyllithium reagent; it is known that a 1: 1 ratio of TMEDA (and other bases) to dialkylmagnes ium reagent is normally found in these _complexes , cf. equation (2-23) , but that o.nly the dimeric form of alkyllithium reagents coordinates (1 :2 complex) with TMEDA . The 52 1:1 RzMg--Lewis base complex, a chelate, has been studied by Zakharkin and indicated to be quite stab le •. 73

Another interesting illustration of the mechanistic difference in the metallations effected by alkyllithium and dialkylmagnes ium reagents arose wi th alkoxide promoted me tallations . It was found that when two equivalents of potassium �-butoxide was added to the non reactive mixture of dialkylmagnesium , substrate (RDE) and Lewis base

(THF or TMEDA) rapid me tallation occurred, presumably via me tal-me tal exchange (2 - 28 , 2 - 29) . An intense dark-red color developed immediately

2K(O�"" Bu) + MgR2 -.2KR + Mg (Ot-Bu) 2 ( 2 -28)

KR + ArH -ArK + RH (2-29) on mixing the reagents at room temperature . When lithium �-butoxide was employed in place of the potassium alkoxide in the mixture , metal lation also occurred, al though much more slowly (refluxing required) and with less intense color formation. Again, metal-metal exchange probably occurred as a first step, followed by me tallation of the substrate (2 -30,2-31) .

Li (O!-Bu) + MgR2 �2LiR + Mg (O!·Bu) 2 (2-30)

LiR + R I -H --.LiR I + R-H (2-31)

In hopes that an exchange between dialkylmagnesium reagents and potassium !-butoxide might allow hydrogenolysia to occur , the me tal-metal exchange reaction was carried out both in the presence and ab sence of Lewi s bases prior to the · admittance of hydro�en gas to the - system. No hydrogenolysis occurred, and . furthermore , the result� ing insoluble produc t subsequently did not promote the. me tallation of

DMR or thiophene . Thus it appears that the metal-me tal exchange must be allowed to take place in the presence of all of the reactant in 74

order for me tallation or occur . Possib le explana tions for th is lack

of me tallation are rather rapid thermal degradation of the resulting

s-butylpotassium (2-32), or me tallation of the Lewis base by the

po tassium alkyl (Z-33) .

K-R -+ KH + olefins (�-32)

� + K-R + M 2N-CH2CH2-NMe2 -+Me 2N-�-CH2-NMe 2 RH ( 2 -33)

Dialkylmagnesium reagents in the absence of Lewi s bases were

found to me tallate a numb er of other substrates , inc luding -picoline ,

dimethyl sulfoxide , and various alkynes such as acetylene , 3-methyl-1-

butyne , and 4-methyl-1-pentyne (cf Tab le XI) . Although me tallation of

alkyne s had previously been carried out with dialkylmagnes ium reagents

in ethereal media , no me tallations (in fact, few react ions in general)

had previously been carried out with these reagents in hydrocarbon

solvents . (This has , of course , been du e mainly to the lack of solu-

bility of the normal C1 -C4 dialky lmagnesium reagents readily obtainable

heretofore in an ether-free state .) It was found that the metallations

of the two substituted alkynes with equimolar quantities of R2Mg pro-

ceeded rapidly at room . temperature with the format ion of soluble me tal- �

lation products (2 -34) . Carbonation of these solutions in each case

yielded a mixture of the two expected produc ts ( 2-35) . Me tallation of

R' -C:CH + R2Mg -+ RMgC:CR ' + RH (2 -34)

( 2 -35)

acetylene (in large excess) by di-.!!,-bUtylmagnes ium in a mixed hy drocarbon TABLE XI

METALLATION .OF ACYCLIC SUBSTRATES BY DIALKYLMAGNESIUM REAGENTS

Reaction Reaction Time Temp . Apparen- t Metallation Yield Reagent Source8 Solvent Substrate Catalyst (hr.) (°C). Product (%)

(!.- Bu) 2Mg I !!.-Hexane- Dimethyl . None Essent. 2 5 Bis-(Methllsulfinyl- Essent. Benzene sulfoxide Instant. methyl)Mg Quant.

c 60% MgC2 + 40% Mg (C2H) �ssent. (!.-Bu) 2Ms II Hexane- Acetylene None Essent. .·,'2.5' 2 cyclohexane Instant. Quant.

(!.-Bu) 2Mg n Hexane- 3-Methyl- None 5-10 25 3-Me-l�b utynylMgR so-1oo cyclohexane 1-butyne- mins. (R•3-M� l-butynyl or ��utyl)d

(!- Bu) 2Mg III Benzene 4-4Methyl- None . 5-10 25 �Me- 1-PentynylMgR -so 1-pentyne mins. (R•4-Me-:-l-pent�yl or !,- butyl)e

a I-Reaction of ·,!.-butyllit,hium with commercial at?-hydrous .MgC12 in hexane-dimethyl etb,er solut�on followed by partial or complete desolvation. ·

II - Reaction of .!.- �utyllithium with synthetic "active" •MgC12 (from (s-:-BuMgCl + Cl2)) in hexane-cyclohexane.

III - Prepared by reaction of t-:-butyllithium in benzene with synthetic "active" MgCl2 (from -....J ,!.-:- BuMgCl + .benzyl chloride) . '-"

1 TABLE XI (CONTINUED)

�roduct formed by reaction of "dimayl"-magnesium with trimethylchlorosilane (TMCS) analyzed by � and NMR and . compared with dimsyllithium-TMCS product. Abe ence of DMS O in product solution. -...... cVPC of deuterolyze� products on triethyl�ine column at -78°C.

�o �- Bu) 2Mg found by NMR analysis.

� showed presenc� of alkynyl- and �- butyl groups in D20 solution of salts obtained on carbonation. · Na2C03 added to aid solubilization.

...... 0\ 77 medium yielded a mixture of magnesium carbide and magnesium acetylide , nei ther of which were hydrocarbon soluble.

Y-Picoline has not. previously been metallated by organomagnesium reagents, alt�u8h it is known to be readily metallated by al kyllithium 110 reagents but only in an ethereal solv�nt . Thus , it is of interest that me tallation by the magnesium reagent takes place readily in hydrocarbon

I solvents at room temperature . A deeply colored red solution is obta��ed du ring the :reaction·. The product was characterized by derivatization with benzoyl chloride to the known solid y- phenacylpyridine (2-37) .

(2-36)

(2-37)

· Grignard reagents are known to form sulfides on reaction with 99b dimethyl sulfoxide (2-38) , whereas alkyllithium reagents me tallate le6. DMSO to form "dimsyl lU:hium" · .(2-3?). In the present study , hydrocarbon- soluble dialkylmagnesium reagents were found to react readily

RMgX + CH3 -�·0 ...... RCH 2-S-CH3 + HOMgX ( 2:-38) �H 3 ; RL:t + CH3-g�=:-+"L:t_qH26t ·+ RH (2-39) with DMSO to give the corresponding di-dimsylmagnes ium reagent , identified by its reac t ion wi th trime thylchlorosilane to give a trime thyl- silyl derivative identical to that obtained with dimsyl lithium ·(2-40) .

(CH2�-<>hMg + 2 (CH3)J-Si-Cl - 2 (CH3 ) 3Si-CH2 -�·0 + MgC1 2 (2-40) eu3 �u3

It seems likely that the anomalous reaction (2-38) obtaine'd wi th the

Grignard reagent is due to prior or simultaneous reaction of DMSO with

MgX2 · 78

E. REACTIONS OF DIALKYLMAGNESIUM REAGENTS WITH BENZOPHENONE

In essentially .all modern studies o� the syst� a Grignard-

ketone complex is proposed to form as a first step (2-41) in the K ketone + Grign�rd complex (2-41) 74 add:ition reaction. For example , Smith has recently presented

evidence for the presence in ether solution of as much as 25 per

cent of the complex between the hindered ketone (2-42) and methyl- ·

.. magnesium bromide when the reactants were mixed in a 1: 1 ratio .

(2-42)

' Reaction in this system was slow and permitted study of the complexa-

tion by ultraviolet techniques. Actually, evidence for such complexes . 101 · was presented as long ago as 1939 by Pfeiffer and Blank. These

authors first advanced the hypothesis that addition is effected by

interaction of the complex with a second molecule of Grignard reagent .

They reported that the mixing of � equ ivalent of EtMgBr with benzo

phenone in ether solution resulted in the formation of a "dirty" white

precipitate which quickly coalesced into an oily insoluble lower ·layer . . .

The oily product , on trea�ent with water , was found to regenerate most

of the benzophenone originally present. When the exper!ment was repeated

with �· (or more) equivalents of EtMgBr, the alcoholate derived from

normal addition was formed. Utilization of an intermediate Grignard-

ketone complex in mechanistic interpretations of their additon reaction

was reviewed in Chapter I. 79

Wh ile essentially no specific experimental work has been reported

. for .dialkylmagnes ium reagent-ketone complexation, it is apparent that

most workers reason that these anal�gous complexes do form and should

serve as addition reaction intermediates . In the present study, branched

chain dialkylmagnesium compounds in hydrocarbon solvents were reacted

in different proportions with benzophenone and the produc ts studied by

NMR and ESR spectroscopy . Evidence for relatively stab le dialkylmag-

nesium reagent-ketone complexes has been found, but the intermediacy

of these complexes in the addition react ion is left open to ques tion.

The findings reported below must be cons idered preliminary and, for

the mo st part, inconclusive , but these findings would seem to make the

further study of these complexes quite promising .

Rapid mixing of one molar equivalent of di -!-butylmagnesium with

two molar equivalents of benzophenone (called hereafter the "1 :2 solu-

tion") gives a pers istent ( 2-3 days) deep brown color wh ich is inter-

preted as indicating the presence of some type of relatively �table

charge transfer complex. NMR examination of this 1:2 solution showed

the presence of benzophenone although its am ount coul d not be accurately

ascertained because of absorbtion ·by the solvent employed, benzene . A

!-butylmagnes ium species was also present , although the !-butyl proton

signals were shifted downfield from their normal position by £!. 0.3

ppm. Addition of another mole equivalent of di-!-butylmagnesium, to _

give a "1:1 solution ," caused the simultaneous .disappearance of the

brown color and the typical benzophenone signal and an increase in

. the !-butrl proton signal . That this !-butyl proton signal was made .·

80 .. up from !,-butyl protons of a magnesium reagent , as well as from a product

carbinol derivative, was determined by NMR analysis after hydrolysis of

the 1: 1 solution. �he hydrolysis of the reaction solution caused loss

of the organometallic !,-butyl signal, bu� retention o� the signal from

the pr�duct carbinol. The product carbinol was , as expected, mostly

the normal addition product.

An ESR spectrum was also obtained with another 1:2, brown

coloFecl benzene s.al�ti..a� Q£ ..

The spectrum which was intense and well-resolved, showed ;at least 44

lines . Both the color and the spectrum disappeared after about one

hour in the sample tube used in the spectrometer.

Rather analogous observations were ma de with di-�-butylmagnesium.

When the secondary reagent in hexane-cyclohexane solution was added to

two molar equivalents of benzophenone in cyclohexane a green, sticky

curdy precipitate formed. (Perhaps th is ins oluble complex .Precipitated

from the 1:2 solution was similar to that ob served by Pfeiffer and 101 Blank.) The green precipitate could be di ssolved in benzene (but

not in ether) with the formation of a green solution which di splayed

an ESR spectrum , not as well resolved as that from the brown 1:2 solu-

tion produced using di -!-butylmagnesium . An NMR spectrum of this

green solu tion showed no trace of the �-butyl methine proton signal

in its normal region for the magnesium reagent , but benzophenon� was

present .

The ESR spectra from th e highly colored 1: 2 sol�tions described

above (brown and green with di -!,- and di-�-butylmagnesium, respectively) 81

have no t been fully analyzed; but most probably the basic structural

unit giving rise to these spectra is that· of a benzophenone radical

anion. An interesting point is that these . spectra show;hyperfine

splitting, despite the fact that the so lvent employed was hydrocarbmn.

These would seem to be the first resolved ESR spectra of ionic speci�s ever obtained in a hydrot=arbon · medium.

From what is known about dialkylmagnesium reagent reactivity

(Chapter I) , there can be little doubt that in the 1:2 reagent : ketone

solutions , ther� is an insufficiency of reagent to cause an addit�on reaction with all of the ketone. That is , when the initial equivalent

I of butyl groups is consumed by an addition (2-43) , the secon� equivalent of' ·

'· i

U.11reacted ketone ·tn the 1:2 solution •• This excess.· ketone ap parently

I then is reduced to a radical anion by the initial adduct, ROMgBu , which

is both a magnesium alkoxide salt and an organomagnesium reagent. Russell, 102 �lanz'en; and Stromm · have pJ'ev.io.us:ly,:demdns't"ated the geni!ta):. aoUd.ty of an alkoxide (RO-) to reduce a single elec tron acceptor (Z) , as represented

in equation (2.:44) . Tentative ly then, radical anions in the 1:2 solu"' ,. . tiona are represented as arising in the manner illustrated in equation

(2-45) .

RO : + z-Ro · + z� (2-44) + ,: ROMgBu + Ph2CO --..(ROMgBu) (Ph2C�) . (2-45)

Just as observed with the brown colored 1:2 solution produced with the. t-butyl reagent , color could be caused to disappear from the .... 82 green solution by addit ion of sufficient di��-butylmagnesium to give a 1:1 solution. However, the rate of addition of �-reagent had to be controlled to produce a SQlutien free. of color. Relatively slow · (drop wise) addition of di-s-butylmagnesium solutiQn to benzophenone solution - resulted in the generation of heat indicating reac tion. A green color

(with ra transient violet cast before thorough mixing) was present throughout the addition, until addition of a molar excess of di-�-butylmagnesium over the 1:1 ratio ca�sed the green color to disappear . At this point reaction of the ketone was apparently complete, i.e. , there was no reduci�le substrate left to accept the electron transfer sug gested in equation (2-45) . After hydrolysis, the product ratio of bentbydrol (reduc tion product) to normal addition product obtained in the case of di-�-butylmagnesium and benzophenone in benzene solution was about 3 :1. Th is can be compared wi th a ratio of only 0.7:1 obtained with the corresponding Gr ignard reagent and benzophenone in diethyl ether .

The most interesting result·in the present . study stemmed from the observation of a transient violet cast when the di-�-butylmagnesium was added slowly to benzophenone. When an excess of the reagent was mixed : rapidly with benzophenone in benzene or hexane-cyclohexane, the - solution became violet colored and the color was stable for an indefinite period�

In other words , ·if the dialkylmagnesium reagent is added slowly to the · ketone, perceptible heat is evolved and a green colored solution (or precipitate) is . formed until a 1:1 molar ratio is achieved , when color disappears and there is no evidence for further reaction , but if the reagent is added rapidly to the ketone so as to give quickly at 83

least ' ca. 2:1 molar ratio of reagent to ketone , a relatively stable violet-colored solution is produced. The following two additional ob servations were made concerning ·the violet solution. (1) It showed an NMR signal from the me thine proton in di-�-butylmagne sium ; no such

NMR signal from the green solution appeared. (2) It showed.an intense .. we ll-resolved ESR spectrum which was different from.that of the green

solution. The spectrum from the violet solution, wh ich is · shown in

Figure 2 , is unquestionably that of the benzophenone radical anion.

Again any interpretation mu st be highly tentative , but the facts concerning the violet solution show clearly that benzophenone , in the form of its radical anion in a hydrocarbon medium , can survive unreacted for an indefinite but long period of time in the presenc e of a large excess of magnesium reagent. A possible interpretation for these facts

is that the large excess of magnesium reagent complexes the ke tone in

such a manner that none is free for additional reaction , and that the species ·so produced is a type of charge transfer complex, as suggested in equation (2-5) . In the presence of limi ted magnesium reagent , or ! Bu2Mg + Ph2c::::i) � (Bu2Mg)+(Ph2CO) (2-5) probably in an ethereal solvent, reversibility of the reagent•ketone complexation (2-5) wou ld be sufficient to always leave ketone free for other reactions .

If the above interpretation is valid , then the idea that the addition reaction requires complexation of a ketone with a magnesium reagent becomes somewhat less attractive. In other words, the above ,.

00 Figure 2. E.S,.R., spectrum-��c�:&&,--(A-but-y�) 2Mg plus benzophenone in hydrocarbon solution .p. (violet color) . 85

interpretation is that complexation ( 2 -5) may be a reversible side

re action , wh ich under some circums tances actually can leave no ke tone

free to be added to.

F. MISCELLANEOUS REACTIONS WITH DIALKYLMAGNESIUM REAGENTS

1. Attemp ted Addition of Dialkylmagnesium Reagents !£Ethy lene

Comparable resul ts to those ob tained in the me tal lation and

hydrogenolysis experiments were ob tained when attempts were made to

add di-�-butylmagnesium to ethylene in a hydrocarbon solvent. No

uptake of ethylene by the reagent occurred in the presence of TMEDA

(alone) or in the presence of potassium 1-butoxide (alone or in combina

tion wi th THF or TMEDA) . Under conditions wh ich caused the absorption

of ethy lene in the presence of �-butyl lithium to be too fast to meas

102 ,104 ure , an equimolar mixture of �- butyllithium and di-�-butylmag

nesium did not ab sorb any significant quantity of ethy l ene . These runs

were made with ethylene at atmospheric pressure ; magnesium alkyls are

known to add to ethylene at elevated temperature and pressure .

2. Attempt ed Reaction of di-A-Butylmagn esium Reagents with Phenyl -·

dime thylchlorosilane in Ethyl Ether

No reaction could be caused to occur between di-�-butylmagnes

ium and phenyldimethylchlorosilane in diethyl ether-hydrocarbon mixtures even after prolonged reflux . The chloros ilane reacts readily with �-butyllithium. 86

l. Effect 2t Di alkyl magnes ium Reagents � t�e Hslgien-Metal ' Interconversion Reaction � Alkyl ltthium Reagents

It was found that dialkylmagnes ium reagents inhib.it th�

halogen-metal interconversion reac tion between alkyllithium · reagents

and �-bromodimethylaniline . When a 2:1 molar mixture of . �-butyllithium

and di-�-butylmagnes ium (in hydrocarbon solution) was treated with 7 �-bromodimethyaniline under the conditions of Gilman Color Test IIa,

only a faintly positive test was obtained. A strongly positive test

was obtained wi th a solution of comparable concentration in �-butyl-

lithium, no magnesium reagent present• · The reactions involved in

obtaining a positive test, shown below, involve first a halogen- metal interconv.ersion •

. �-C4H9Li + {CHJ} 2N-C6H4Br -..(CH3) 2N-C6Hi.Li + C4H10 (2-46)

{CH3) 2NC6H4Li + (C6Hs) 2co _... (CHJ) 2NC6114-CO{Li) (C6Hs) 2 (2-47) I

(red) (2-48)

Proof that inhibition of the halogen-metal interconversion (2-46) was occurring when the mixture of the two organome tallics was used was obtained by carbonation of appropriate reaction mixtures and NMR

examination of D20 solutions of the carboxylic acid salts so obtained.

When �-butyllithium alone was allowed to reac t with �-bromodimethyl- ·

aniline , carbonation indicated that all of the alkyllithium exchanged

to �- dimethylaminophenyllithium. In the case of the mixed lithium- magnesium alkyl , only an ins ignificant amount of exchange occurred ..

and mainly carbonation products from the unchanged �-butyl reagents were ob tained. 87 These results� those from hydrogenolysis experiments, from solubility studies and from NMR spectral measurements , all indicate that mixed magnesium- lithium alkyls cons titute a new group of complex struc tures. They would seem to be new hydrocarbon-soluble struc tures in which one electron-defic ient member of the oligomeric unit has simply been replaced by a different organometallic molecule possessing like properties. The minimum molar ratio of alkyllithium to dialkylmagnesium reagent in such complexes appears to be 1:2, but it is not known what the upper limit might be . Indeed,. it may vary widely , and should be the subject of further research. The halogen-metal interconversion react ion might be a useful · diagnostic· test to determine this limit. ·

G. HOMOLOGIZATION OF ALKYLLITHIUM REAGENTS

During an attempted preparation of di-�-butylmagnesium from exchange of �-bu�yllithium with·MgCl2 in the presence of dimethyl ether , it was observed that some of the �-butyllithium reac ted with the dimethy l ether. A homologization product, 2-methylbutyllithium was formed, in addition to the exchange product. The homologization product is also formed in the absence of MgC1 • Its formation can be _ 2 formally represented (2-49 � 2-50) as involving a carbene insertion

CH2 : + CH3-CHz-ys-cH3 --+CH3CH2-�H-CH2-Li (2-50) Li CH3 88. process . One can conc eive of this reaction as proceeding by me tallation

of DME to yield an intermediate "carbenoid, " e,g. , CH30CH2Li, wh ich

causes the insertion. It is known that CH3oCH2MgCl gives carbene-type reactions . 103 The overall. stoichiometry of this insertion process can

be represented as consuming two solvated dimers and one free DME mole-

2 2 cule ( �51) �· For :the purpose of discussion of the equation ( - 51) , .

the DME whi ch is free or unc omp lexed by dimeric alkyllithium reagent

is called "effec tive DME ," DME eff• 2 2 (RLi) 2 :DME + DME ---+{RCH2Li)2 :DME+ 2Li0Me + 2RH ( -51) !-Butyllithium also reacts with dimethy l ether and is homologized

to produce neopentyllithium. The insertion produc ts have been identified by NMR al an ysis and by conversion to the lithium salts of the expected carboxylic ac ids on carbonation. The experimental procedure for this homologization constitutes a new , rapid means of converting secondary and tertiary alkyllithium compounds to their corresponding branched- chain normal homologues containing one more carbon atom . Yields are essentially theoretical ; two mo les of RLi produce one of RCH2Li . The kinetics of the homologization reaction were examined wi th

�- butyllithium in hydrocarbon me dia using two types of experimental conditions: runs involving only the reactants shown in equation (2 -51), designated as "ordinary ru'!ls ," atrd runs having tetramethylethy lene diamine present designated as "TMEDA-catalyzed runs ."111

1. Two Ordinary �

The first two kinetic runs were carried out at 40 0 in cyc lohexane with just reactants present . In one , only the amount of DME 89 .

corresponding to the stoichiometric amount depicted in equation (2-,�)

' · was employed. In the other run, an excess of DME was employed, wh ich

i.nsured that the effective . DME present would remain 'essentially con-

stant throughout the run . The reactions were allowed to proceed to

over 50 per cent completion and in each case samples were withdrawn

periodically , reacted with excess trimethylchloros ilane and the pro-

duct solutions assayed for s-butyl- and 2-methylbutyltrimethylsilane -

by gas chromatography . The ' fraction of unreacted !-butyllithium (£).

'was treated as the reaction variable and various· plots involving c

versus time (1) were made for each run, including c !.!• t, log c .!!.• 2 . t:. 1/.c -vs. t, and l/c -vs. t. Flgure 4 depicts the results from the .

s·toichiometric DME run and Figure 6 , the results from the excess DME

rul,l. (cf. also Figures 3 and 5 for 0-5°C data) .

With a stoichiometric amount of DME , a plot of l/c2 !!• t

appeared to fit the data best (Figure 4) , wi th excess DME, a plot of

1/c .!!.· t fit the data best (Figure 6). These observations are con-

sis tent with the kinetic order of the reaction� being, . ·somewhat sur-

prisingly, exactly the same as the stoichiometry represented in

.equation (2f51) . That is, .the process is third order overall, second

order in solvated dimer reagent , and first order in the free or effec- ·

tive DME as indicated in equation (2-5'2) . 2 [(RLi) :DMEJ [DME] 2-52) rate • kr 2 eff (

In order to calcula te the rate constant (!r) for the run with excess DME , equation (2-52) can be rewritten as in (2-53) in wh ich the

( 2-53) 90

3.0'

2 .0 1 �

1.0

Time

Figure 3. Homologization of �-butyllithium wi th stoichiometric 2 dime thyl ether at 0°C.l/(fraction unreac ted �-butyllithium) versus time (minutes) . 91

6 1 7

200 400 600

Time

Figure 4. Homo1ogization of �-buty11ithium with stoichiometric dimethyl ether at 40°C; 1/(fraction unreac ted �-buty11ithium) 2 versus time (m inutes) . 92

_L c

120 240 360

Time

Figure 5. Homologization of �-butyllithium wi th excess dimethyl ether ether at 5°C; !/(fraction unreacted s-butyllithium) versus time (minutes) . 93

3 2

...L c 16

1 20 240 360

Time

Figure 6. Homologization of _!-butyllith ium with excess dimethyl ether at 40°C ; !/(frac t ion unreacted -s-butyllithium) versus time (minutes) . 94

initial solvated dimer concentration is given as [RLiJ i/2, and the

effective DME concentration is expressed as [DME]xs ' .because it is in such excess over that required stoichiametrically as to be a reac tion

constant . Under the circumstances the only variable is c. On inte-

grating equation (2-53p , equation (2-54) is obtained, wh ich shows that

a value of kr in the run with excess DME can be ob tained using the

slope of the 1/c �· t plot in Figure 6. Doing this, one ob tains a

1 -1 40°. value for kr of 0.134 1. mo le- min. at [RLi] /

= + 1/c kr [DMElxs ( 4 ) t constant (2-54) The rate constant for the reaction with stoichiometric DME can be derived as follows . In this case, [DME]eff is always just one-half

the unreacted solvated dimer concentration, so the rate expression may

be written as in (2-SS} . Integration gives 'equation (2-56) , wi th

< [RLi] i 3 rate = kr ( 2 r c (2-55) 3 [RLi] + .1 k constant (2-5 6) --z = r c ( 16 ) t 2 which, and the slope obtained from the l/c vs . t plot (40° data, 2 -2 -1 Figure 4) , one obtains a value for kJ:. of 0 •. 151 1 mole min .

' The agreement between values of Kr (0 . 14±. 01) ob tained from

these two runs is remarkab ly good. The conclus ion is that in these runs the order of the homologization reac tion fo llows the stoichiome·t ry

I shown in equation (2-51) , i.e. , the insertion process is third order.

Two solvated dimer units as well as the free base , (DME)eff, are involved 'before the rate determining step. This result is consistent . 95�oo I with the concept that the free base is a catalyst, a concept previously 100 suggested by Screttas � for reactions -of lithium res.gents ' with other

substrates . · ·

The involvement of the two solvated dimers does not seem to be

consistent with formation of .any free carbene, or even a carbenoid.

From the stoichiome try and kinetics, it seems possible that one sol-

vated dimer unit is the "reagent" and is ac t ing on a second as the

"substrate ," causing the second to undergo the insertion reaction . An

attempt to represent this general idea is shown below. Quite possibly

the primary action of the one solvated dimer on the other is charge

transfer and this transfer process i• catalyzed by the free DME .

reagent catalyst substrate 2 (R2Li2)DME + DME + (R2Li2)DME ( -57) � 2 2 RH + LiOCH3 + (RCH2Li)2IME products

2 . !2!!!:, TMEDA-Catalyzed !!:m!.,

Two kinetic runs were carried out at low temperatures with equal

catalytic amo�nts of TMEDA but with differing amounts of excess DME.

Two other runs were carried out at 0°C with differing amounts of TMEDA

and the DME present in stoichiometric amount . The data from these

catalyzed runs were plotted as c !!.· t, 1/c .!!.• t,,,,l/c2 !!.· t, and

log c !!.· t. The best fits for the runs with excess DME were obtained ,. ,

by plotting log c ll· t and are shown in Figure 7. The best fit for the

runs with stoichiometric DME were also obtained by plotting log c ll·

t and are sh0wn in Figure 8. In other words , the . reaction is no longer

third order ; all of the TMEDA- catalyzed runs appeared to be first order

reactions . -...... The result that the overall reaction appears as first order in

the presence of catalytic amounts of TMEDA is analogous to results lOOb. . it.\ ...: . on catalysiS o.f found by Screttas . ·• his ·sj:udy. the hydrogenolysiS of

RLi reagents by various bases . He observed that with most monofunc-

tional bases , such as tertiary amines and ethers, �ydra,geno�ysis· occurred

at a moderate rate and was kinet�cally dependent (first order) on RLi

reagent concentrations even with excess base. With difunctional bases

such as TMEDA as catalysts , RLi hydrogenolysis rates were too high to

study with excess base. Measurable rates ob tained by Screttas100b with

low. l'MEDA concentrations showed no ki�etic dependence on the RL i.reagent

concentration, although the rates obtained did show kinetic. dependence . 104 · tha t' on catalyst (base) concentration. Selvidge .r.lat·er showed :... the'� TMIU>A-

catalyzed hydroge�olysis does show a kinetic dependence also on the con

centration of substrate , hydrogen. Thus, it is not surptising that the

TMEDA-catalyzed homologization reaction, whe�ein solvated dimer acts as

both the substrate and the RLi reagent, should be only first order in

solvated dimer.

The loss of two kinetic orders observed (in going 'frem ' ordi�;ry

runs with stoichiometric DME to TMEDA-catalyzed runs) is due to free DME ,

as we ll as one solvated dimer unit, no longer playing kinetic roles in 97

1.0

0.8

. 0.6

0.4

0.2 c 0.1 o:o8

0.06

0.04

0.02

40 80 120 Time

Figure 7. · Homologization of .!,-b utyllithium with excess . dimethyl ether catalyzed by TMEDA (0.0134M) ; fraction unreacted .!,-butyllithium versus time (minutes). 98

9 8 7

6 ---o ETMEDA = O.OlOM 4 ----o-

3 TMEDA = 0.033M

150 300 400

Time

Figure 8. Homologization of �-butyllithium with stoichiometric 0 dime thyl ether catalyzed by TMEDA at 0 C; fraction unreacted �-butyllithium versus time (minutes) . 99 the reaction. That there is loss of kinetic dependence on the free DME concentration in the TMEDA-catalyzed runs is corrobo�_ted by the fact

that there ·was no change in the apparent order of the reaction in the presence or absence of excess DME , although there was such a change in ordinary runs . This loss almost surely is a consequence of the TMEDA being ·a much better catalyst than DME . From the work of Settle 105 · on steric effects in RLi solvation, it is safe to assume that .the dimeric

RLi units present would be preferentially solvated by DME , · rather than

TMEDA: The point is, it is free TMEDA concentration wh ich promotes catalyzed runs , and even in the ordinary runs , free DME was just a base acting as catalyst and was not the substrate. The substrate in all runs was dimer solvated with DME , wh ich was caused to react with the reagent , another solvated dimer, by one base or the other acting as a catalyst.

It follows that the overall rate in the TMEDA-catalyzed runs was probably a two-term func tion, equation (2-58) , the first of wh ich was 2 rate· = (krliJMl [IME] eff [ �2Li2i-JME] ·�. +. (kr.)�}c[JI'MEDAle·ff[R2'1li2:DME) (2-58) · - ·�.. insignificant in magnitude to the second. Hence the integrated func - tion could be written as in equation (2-59) . From the log c �· t

(2-59) plots, slopes yielded the rate constants· for TMEDA�atal,Ze.t runs::wh'ich

•t�· !?l'\Q.� Jn �a:bJ�e X:� Fo·r · CQ.'I't:�bora'tion't hat · the. ·prqper rate" function 100

3. ·-Butyl lithium �

One kinetic run was made employing £-butyllithium and excess DME , but no TMEDA. The best fit for the · s-butyllithium-excess DME reaction

data was a first order plot (log c :!!· t) shown in Figure 9, which is in

contrast with the fits of a second order plot for �-butyllithium-excess

DME runs .

Again, analogous contrasting results were found in the hydrogeno

lysis study. Other RLi reagents (i.e. , n- and �- reagents) and monofunc 

tional bases gave data showing kinetic dependency on the RLi concentration, but £-butyllithium was k�netically independent of that reagent even with

the monofunctional basic catalyst. Thus , £-butyllithium with DME ac ts

like �-butyllithium with TMEDA; there is a washing ou t of one kinetic

order with respect to solvated dimer. Perhaps this is due to the fact

that s-butyllithium is a better electrQn source than is �-butyllithium.

The bright yellow color, clearly ob servab le du ring the !-butyl

lithium-DME reaction, (but ab sent or only faint in the �-buty llithium-DME reaction) is thought to be evidence for a charge transfer complex or a

radical ion intermediate and thus to support some electron-transfer pro

cess. This coloration disappears when the reac tion is essential ly complete.

An ESR spectrum (Figure .lO) of the yellow solution showed a weak , but

definite signal .

The rate func tion for this !-butyllithium reac tion would be that

in equation (2-60) , where (DME) eff is again essentially a reaction constant

(2-60) .. 101

0 .. 1

0.0

o.o

0.0

Time

Figure 9. Homologization of �-butyllithium with excess dimethyl ether at 0°C; fraction unreacted �-butyllithium versus time (minutes) . 102

. � ...... 103 '>....

equal at any time to ene -half tne unreacted solvated dimer concentra-

tion plus the - excess DME concentration, t.e., that above .stoichiometric .

' 2 · The kr value fo� 1.5 ·x 10- , for the !_- butyllithiuui run , and those

for other runs , are shown in Table XII . From estimates of the experi-

mental error, the rate constant s shown in Table XII are considered to

be no more accurate than about t 10 per cent of each value shown .

4 • SU11D118ry

Comparhc:m of the rate constant s shown in Table· XII does no. t .constitute a part;ic\llar�y \lseful comparison of homologization .kinetics under

the various conditions emp loyed. The primary reason for this is the

fac t that these values are mixtures of second and third order constants.

· Even th is fac t that the homologization kinetic . order changes in going

from one experimental condition to another has not been really rationa-

lized with a reaction mechanism, although analogous changes in another

reaction (hydrogenolysis) have been cited, along with ·the idea that the

kinetic role of base (its catalytic activity) is played as a charge

transfer agent . In this connection it has recently been pointed out

that while the general concept of base support of charge transfer has

obvious utility in rati'onalizing diverse phe�omena , eluc idation of

4ettiled steps in reaction mechanisms may provt . difficult.� · For the

present , our eluc idation of the homologization reaction cons ists mainly

of the two kinetic expressions written below.

rate = kr[R2Li2:DME] [base] [R2Li2:DME] (2-61) reaJent catalyst substrate

rate ... kr[base] [R2Li:DME] catalyst substrate 104

TABLE XII

KINETIC DATA-REACTION OF s -BUTYLLITHIUM WITH DIMETHYL ETHER

Run No . Temp . °C [THEDA ) [DMB1xs (RLi}t kr

1. 40 1.1 1.3 0.13b

2.a 40 1.24 0.15c

b 3. 5 1'. 1 1.31 7.4xlo-2 a c 4. 0 1.24 9.6xlo-3 d 5. 2.5 0.0134 1.0 1.31 o.8 d 6. 4 0.0134 1.5 1.31 1.4 d 7. 0 0.010 0.80 l.l d 8. 0 0.033 0.59 0,9

f - e 9. -7 - 0.8 1. 75 1.5xlo-2

a Arrenhius activation energy calculated from data on runs 2. and 4. is 2.9 kcal/mole .

b (�lope)(4) Function for k (calc 'd) = ; units = liter2 r 2 2 -1 mole- min (third order) . [DME] XS [RLi] i c (slo2e)�l6) 2 Func tion for k r (calc 'd) = units = liter2 mole- 1 3 min- (third order) . [RLi]f d (slo2e)�2) Function for k r (calc 'd) = units • liter inole-1 [TMEDA) [RLi) min-1 (second order) . i e (slo2e) (2) Func�ion for k r(calc 'd) • units • liter mole-1 [DME] 8 [RLi ] min-1 (second order) . x f f �-Butyllithium used in run instead of �-butyllithium. 105 The substrate irt the hom ologization reaction is the same species

as the reagent. With !,•butyll ithium as reagent and DME · as catalyst,

equation (2- 61) is followed; there is kinetic dependence an both reagentr· ,I and substrate. With the !,• reagent and TMEDA . as catalyst, equation (2-62)

is followed; there is no kinetic dependence on reagent. The latter is

true with !-butyllithium even with DME as catalyst. The greater reac-

tivity, i.e., loss of kinetic dependence , with !•butyll ithium and with

.t·butyll ithium-TMEDA , quite likely results from greater fac il ity in

charge transfer.

Values wh ich do constitute a useful comparison of homologizati�n·

kinetics under the various conditions are shown in Table XIII . The

numbered entries there correspond to thoee in Table XII, but the rate

values shown are calculated ab solute initial rates (-dc/dt) wh ich woul d

have ob�ained had each of those runs been initiated with. RLi • 1M.

For example, from these values one comparison available (runs 4 . and 7,

Table XIII) , is that addition of j.ust 10-2 mole equivalents of THEDA.

to a stoichiometric mixture of -s-butyllithium and IJoiE causes a ten-fold

increase in their reaction rate. 106

TABLE XI II

INITIAL HOMOLOGIZATIQN RATES OF 1M RLi WITH DME

Initial Relative a [ b Run No . Temp . [TMEDA ] DME l xs Rates Rate

0 1. 0: 60 4 1.1 0.036

2. 40° 0. 009 15 0 3. 5 1. 1 0.020 35

o 4 . o 0.0006 1 0 5. 2 0.01� 1.0 0.005 8 0 6. 4 0.013 1.5 0.009 15

o 7. o 0.010 0.006 10

o 8. o 0.0\33 0. 015 25 0 9 •. -7 0. 5 0.006 10

a All runs with �- butyllithium except Run 9. with t-butyllithium•.

b Values calculated from rate constants in Table IX. CHAPTER III

EXPERIMENTAL

A. INTRODUCTORY INFORMATION

. 1. Reagents

.u-Butyllithium in .n-hexane and benzene solutions and !,-but.yl

--l.ithium irt ,!l-hex_��e and cyclohexane solutions were obtained from

Foote Mineral Company . S,-Butyllithium in .u-pentane and methyl-

lithium in ether were ob ta ined from Lithium Corporation of America�

Lithium (containing 2 per cent sodium ) as a 30 weight per cent dis-

persian in mineral oil, and as 1/8 inch wire was also ob tained from

Lithium Corporation of America. !- Butyl- , isopropyl- , and cyclohexyl-

magnesium chloride in ether solution were ob tained from Matheson,

Coleman , and Bell; phenylmagnesium chloride in tetrahydrofuran was

ob tained from Mallinckrodt Chemical Company . Anhydrous magnesium

chloride in powder form was obtained from Alfa Inorganic s, Inc . and � in lump form from the Dow Chemical Corporation. Anhydrous triethylene

diamine (TED) , m. p. 158-160° , and methyltriethy lenediamine , a Uquid-,

were obtained from Houdry Process and Chemical Company . N,N,N' ,N'-

tetramethylethylenediamine (THEDA) was ob tained from Rohm and Haas

Company . The THEDA was distilled from lithium aluminum hydride and

stored over Linde 4A mo lecular sieves; otherwise, the reagents indi-

cated above were used without refinement.

107 108 Other critical solvents and reagents (benzene , toluene, cyclohexane , thiophene , resorcinol dimethyl ether , Y-picoline , etc.) were reagent grade (e.g. , Baker 's Analyzed Reagent , Mallinckrodt Analytical

Reagent, Eas�man White Label) and generally were dried with activated alumina and stored over Linde Molecular Sieves �-A (MS -4A) . Benzene and cyclohexane used for desolvation of dialkylmagnesium reagents and for active alky l analyses were specially treated; these solvents were refluxed for at least 8 hours over lump barium oxide and then distilled into bot tles (previously dried at 125°) containing Linde MS -4A sieves and. a few pieces of sodium me tal . Ether and tetrahydrofuran were refluxed over lithium aluminum hydride and distilled into amber bottles containing

Linde MS-4A sieves and sodium.

Nitrogen was a high purity grade obtained from Welding Gas Products

Company , was passed through two drying columns , the first containing indicating drierite and the second, Linde MS-4A Molecular Sieves . Dry hydrogen gas , also from Welding Gas Products Company , was passed through a wash-bottle containing conc entrated sulfuric acid. Ethylene and dimethyl ether were ob tained from the Matheson Gas Company ; each was indicated to be at least 99.5 per cent pure and was used without further purification.

2. Ins trumentation and Analyt ical Techniques Magnesium analyses were carried out using a conventional complexo metric titration involving an ethylenediamine tetraacetic acid di-sodium salt solution as titrant and "calmagite" ob tained from the G. F. Smith Chemical Company as indicator. Chloride ion was determined by the standard Mohr procedure. Total alkalinity analyses were carried out 109

by adding a known amount of standard aqueous acid (in excess) and back

titrating dropwise (with vigorous shaking) with standard aqueous base .

Each of these titrations was done on hydrolyzed samples of organometallic

solutions .

Active alkyl analyses were done by direct titration of the organo

me tallic solution samples using the procedure of Watson and Eastham� 98b

Gilman Color Test I, which is sensitive to the presence of both

organolithium and organomagnes ium compounds , involved the addition of

0.5-1.0 ml . of the organometal lic solution to be tested to an equal

vo lume of 1 per cent Michler 's ketone in dry benzene , followed by

hydrolysis with 1 ml . of water and addition of several drops of a 0.2

per cent solut ion of iodine in glac ial acetic acid. A greenish-blue

color was cons idered positive indication of the presence of C-Li or

C-Mg containing compounds . Gilman Color Test Ila was used to differ

entiate between alkyllithium and dialkylmagnes ium compounds and involved

the addition of a 0.5-1 .0 ml . of the organome tallic solution to an equal

volume of a 15 per cent solution of �-bromodimethylaniline in dry benzene,

followed by 1.0 ml . of a 15 per cent solution of benzophenone in dry benzene . After a few seconds , the mixture was hydrolyzed with wa ter

and acidified with hydrochloric acid. A red coloration indicated the

presence of an alkyllithium compound. Since it was found in this work

that dialkylmagnesium compounds interfere in this test, the absence of alkyllithium reagent as de termined by th is test was in mos t cases veri

fied by magnesium and total base analyses . 110

NMR spectra were obtained with the Varian A-60 Spectrometer .

Vapor phase chromatographic analyses (VPC) were performed with the

Perk in-Elmer Model 154-D Vapor Fractometer . Details of typical VPC

analytical procedures are given on page 1�9. ESR spectra were obtaine d

with the Varian 4500 with 100 KC Modulation (frequency adapted) . Me lting

points were determined with a "Mel-Temp" apparatus and are reported

uncorrected.

B. PREPARATION OF DIALKYLMAGNESIUM REAGENTS FROM GRIGNARD

AND ORGANOLITHIUM REAGENTS

1. Preparation .Qi Q!.-!:- butylmagnesium·.

A we ight of 2.43 g. (0 .1 g. atoms ) of magnesium powder (40 mesh)

and a few crystals of iodine were placed in a 500 ml . round-bottom

4-neck flask equipped with a mechanical (Hershberg) stirrer , reflux

condenser , thermometer , and 100-ml . graduated dropping funnel . The

flask was heated to vaporize the iodine and then allowed to cool .

Thre.e to four milliliters of a ·solution· of .9��3 .. g . (0.1 mo le) ()f _t- butyl

chloride in 25 ml . of anhydrous ethyl ether was added to the flask in

one portion . Reaction commenced almost immediately. After adding 25 ml . of ether to the reaction mixture in the flask, addition of the

rema inder of the hal ide solution was continued with refluxing and stir

ring. After addition of 10 ml . of the hal ide solution, the rema inder

of the halide solution was diluted with 20 ml . of additional ether and

addition continued over a period of 1 hour with vigorous stirring. A

3-ml . aliquot of the clear solution was analyzed for total alkalinity 111

and found to contain 0.94 meq . per ml . of solution. The total 59 ml .

of solution thus contained a total of �· 56 meq . of �-butylmagnesium

chloride , i.e. , the yield was 56 per cent .

By means of a syringe , 35 ml . of a 1.44N solution of t-butyl

lithium in n-pentane was added rapidly to the stirred solution of Grignard

reagent . Some heat was evolved and formation of a light gray preci

pitate of lithium chloride was noted. After it was stirred for a few

minutes , the mixture was allowed to settle . Aliquots of the clear

supernatant solution were withdrawn for chloride and total base analyses ,

which showed 0.50 meq . , of chloride and 1.47 meq . of total base per ml .

of solution ._ By means of a syringe , an additional 20 ml . of 1.44N t-

.butyllithium solution was added rapidly to the stirred mixture and stir

ring was continued for another hour . After allowing the mix ture to

settle, analyses of aliquots from the clear colorless supernatant solu

tion showed 0.016 meq. of chloride , 0.016 meq. of lithium , 1.62 meq. of

magnesium , 1.62 meq. of active alkyl , and 1.63 meq . of total base per ml .

of solution. The yield of di-t-butylmagnesium solution thus obtained

was essentially quantitative , based on �-butylmagnesium chloride.

Th is solution could not be freed of base (ethy l ether) by codis

tillation with added hexane solvent at atmospheric pressure , apparently

due to facile decomposition to MgH2 and butene-1. It could not be fre ed

of base at ordinary temperatures by vacuum distillation at 0.1 mm. pressure .

2. Preparation of Methyl isobutylmagnesium

A solution of Grignard reagent was prepared from 0.2 g. atoms of

magnes ium (turnings) and 0.2 mole of isobutyl chloride in 200 ml . of 112 anhy drous ethy l ether (92 per cent yield) . To this solution was added as rap idly as possible with stirring , 200 ml . of a 0.92N solution of methyllithium in ethyl ether . The mixture was stirred for 2 hours and

then allowed to settle overnight. The clear supernatant solution

(analysis of which showed a ratio of Mg :Li:Cl of 16: 1:3) was distilled and the ether solvent replaced with benzene . The residual solution was centrifuged and the clear viscous supernatant solution on analysis showe d a 1.45N total base , a I.41N active alkyl and less than 0.01 in chlorine content . . Both VPC and NMR indicated an excess of methyl over isobutyl groups , �· 1.2-1.3 methyls to one isobut��·

3. Preparation of Me thy l -,.!- Butylmagnesium

To a volume of 77 ml . of 2.20N ,.!- butylmagnesium chloride in ether , prepared routinely , was added 134 ml . of 1.23N methyllithium solution in ether . Some heat was generated and a precipitate formed.

After stirring overnight, the mixture was allowed to settle . The clear , colorless supernatant solution was analyzed and found to be

1.64 N in total base and 0.057N �in chloride ion. NMR analysis of

the solution showed the ratio of .!� butyl to me thyl groups in the organometallic to be one .

4. Preparation of �-Buty lphenylmagnesium

To 50 ml . of a 1.44N solution of ,!-butyllithium in .!!,-pentane

was added as rapidly as pos sible and with vig' orous stirring, 29 ml . of a 2.4-2.5N solution of phenylmagnesium chloride in THF . Boiling of

the pentane occurred, bu t the reac tion subsided quickly after all of 113 the PhMgCl had been added. After it was stirred for 2 hours , the mixture . was allowed to settle, yielding two liquid layers besides the wh ite solid

LiCl. After distillation to remove most of the E- pentane , only one clear liquid layer remained. NMR analysis of this layer showed the ratio of

£-butyl to phenyl groups in the organometallic to be 1.1.

5. Preparation 2f Diisobutylmagnes ium

To a solution of Grignard reagent prepared from 2.43 g. {0.1 g. atom) of magnesium turnings and 9.3 g. {0.1 mole) of isobutyl chloride in 100 ml . of ethyl ether {90 per cent yield) was added 90 ml . of an

0.96N solution of isobutyllithium in ethyl ether prepared from 1.4 g.

{0.2 g. atom) of lithium me tal wire (con taining 0.8 per cent sodium) and 9.3 g. of isobutyl chloride in 100 ml . of ethyl ether (87 per cent yield) . The mixture was refluxed for 5-10 minutes and was allowed to stir for 5-10 minutes longer and then to settle. Analysis of the supernatant solution showed 0.84N · total base , 0.02N chloride , 0.88N

Mg , and 0.02N Li. The clear supernatant solution was transferred to

. a distillation unit, and the ethyi ether was distilled wh ile cyclohexane was added until the ether concentration in the distillate was

20-30 ppm. A wh ite precipitate formed during the distillation. The mixture was allowed to cool and the wh ite solid was filtered onto a sintered glass {M) filter funnel under nitrogen. The solid was washed with dry E-pentane , blown dry with argon, and transferred to a we ighed vial { fitted with a septum) in a dry box. Analyses of the solid for 114 total base and ethyl ether (VPC) content showed it to possess a neutral equivalent of 66 (theory 69) and no ethyl ether. It had a solubility in benzene of 0.135 meq ./ml .

6. · Preparation of Di -�·butylmagne sium from Simultaneous Preparation of

�-Butyl lith ium and- �· Butylmagne sium Chloride

A crystal of iodine and 5 ml . of a solution of 17.5 (0.19 moles) of n-butyl chloride in 25 ml . of ether were added to 150 ml . of ether and 2 .4 g. (0.1 g. atom) of magnesium (turnings) in a 500 ml . 3-neck flask equipped wi th reflux condenser , mechanical stirrer, and dropping funnel. After reaction began, a weight of 4.7 g. of lithium dispersion

(30 weight per cent in mineral oil) containing 1.4 g. lithium metal (0.2 g. atom) was added. There was a slight heat of reaction evident on mix ing the reagents. The remainder of the -n-butyl chloride solution was added over a 1 hour period, while the mixture was refluxed, and the mix ture was stirred for 15-20 minutes longer. It was then cooled to room temperature and al lowed to settle. Analysis of the supernatant solution showe d 1.04N total base, 0.39N Li , 0.60N Mg , and 0.02N Cl content. The yield of di -n-butylmagnesium was 61 per cent (based on - n·butyl chloride) .

C. PREPARATION OF DIALKYLMAGNESIUM REAGENTS FROM ORGANOLITHIUM

REAGNETS AND ANHYDROUS MAGNESIUM CHLORIDE

1. Preparation of Di -&·butylmagnesium

a. �dimethyl ether-solvated � product .MgC 12• The magnes ium chloride used in the following preparation was produced as a solid by product in the direct reac tion of isopropyl chloride and magnesium turnings 115 in a benzene solution containing dimethyl ether (DME) following the procedure given on page 130 for the direct preparation of diisopropyl- magnesium .

To a stirred slurry of about 24 g. (0 . 25 moles) of magnes ium chloride in 125 ml . of benzene was added 150 ml . of a 1.3N �-butyl - lithium (0.2 mo les) solution in cyc lohexane . The temperature of the

° ° mixture rose to 3? C and then slowly dropped to below 30 . After further stirring overnight , Gilman Color Test IIa on the supernatant solution was negative. NMR analysis showed a ratio of 1.8 mo les of

DME to 1 mo le of di-�-butylmagnesium. Ano ther volume of 77· ml . of 1.3N

�-butyllithium (0 .1 moles) solution was added to the stirred react ion mixture and after several hours, Color Test IIa was again negative . A third charge of 77 ml . of �-butyllithium solution (0 .1 mo le) was added and after stirring for �· 2 hours , Color Te st IIa was negative. Decants- tion yielded 365 ml . of a 0.98N (0 .36 mo les, 90 per cent yield) solut ion of di-�-butylmagnesium . Analy sis of the sol,ution for concentrations of total base , ac tive alkyl and ma gnes ium content showe d the concentrations ·- .. . to be equivalent to each other within experimental error :' Essentially no chlorine (<

b. From ether-desolvated �-produc t MgC 12 . The solids in the slurry from the desolvation experime nt described on page 127 were washed several times with benzene , wh ich left the MgCl2 free of ethy l ether .

These solids were then successfully reacted with �-butyllithium in cyc lohexane exactly as just described above . to give directly a solution of di-�-butylmagnesium free of ether. 116

c. From ether-desolvated MgC 12 �-produc t !rQm reac tion £fA

Grignard reage nt � chlorine . A Grignard reagent was prepared from

24 .3 g. (1.0 g. atom) of magnesium turnings and 92.5 g. (1 .0 mole) of:·,!:-buty1 chiotide in 500 ml of anhydrous ethy l ether. Chlorine gas , diluted with argon, wa s bubbled into the vigorously stirred solution for 3 hours. During .the passage of chlorine , intermi ttent bright flashes of light appeared (chemiluminescence) every one to two seconds .

Toward the latter part of the reaction , a grey-black vapor was noted in the flask and condenser. Chemi luminescenc e has been previously observed on reaction of Grignard reagents with oxygen. 7

The solvent refluxed spontaneous ly, and a th ick, porridge-like mass formed in the flask. Analysis of the supernatant solution for· total base showed th at a 70 per cent consumption of .!-butylmagnesium chloride had occurred. After 300 ml . of benzene was added to the flask, mo st of the ether was removed by distillation . The residual mixture was filtered and the solids washed three times with benzene on a sintered glass funne l. The salts were transferred as a slurry in benzene to a distilla tion unit and the ether was distilled with the aid of benzene and then cyclohexane . Distillation was continued until a hydrolyzed sample of the slurry showed the presence of less than 1 mole per cent ethyl ether compared to MgCl 2 • Fifty milliliters of the slurry wa s reac ted with 50 ml . of 1.3N .!-butyllithium in cyc le hexane . Heat was noted and after 2 minutes , Color Test IIa was nega tive . Al l properties (NMR , VPC and Cl, Mg , Li, and total baae contents) of the resultant solution were those· of di-,!�butylmagne siu111 free of . ethereal solvation. 117

d. � dimethyl ether-solvated commercial anhydr ous MgC12-.

Two hundred ninety mi lliliters of a clear filtered solution was obtained from a reaction of 360 ml •. of l.ON �-butyllithium in !:hexane with 20 .8 g. (0.22 moles)· of powdered anhydrous MgCl2 and 20 g. (0.44 moles) of dime thyl ether. This solution , wh ich was 0.86N in total base , was placed in a 500-ml . 3-necked round-bottom flask equipped with an 18- inch Vigreux column attached to. a distillation head, condenser , and distillate receiver. After distillation of 200 ml . of solvent , an equal volume of fresh hexane was added to the flask. Thereafter , five

40-ml . fractions of solvent were distilled consecutively . The per cent

DME in these five cuts were , respectively , 1.8, 1.1, 0.83, 0.65 and

0. 12, as determined by VPC . A one mil lil iter sample of the residual solution (76 ml. ) in the distilling flask was solvolyzed by adding it slowly (via a syringe) to 3 ml . of isopropyl alcohol at 0° contained in a centrifuge tube fitted with a rubber septum . After centrifugation, the clear supernatant solution was analyze.d by VPC for DME and n-butane . This analysis indicated a mo lar ratio of n-butane :DME of

12:1 or an initial mo lar ratio of di-�-butyl magnesium to DME of 6:1.

Ten milliliters of the above residual solution was further desolvated by distillation from a smaller flask attached to a Vigreux column and fractionating head. During the distillation , the hexane solvent was replaced with benzene . After distillation of about 50 ml . of solvent , some decomposition or cracking of the product began to be apparent and distillation was discontinued. NMR·analysis of the residual solution showed the. absence of DME . 118

e. � ethyl ether-activated, desolvated commercial anhydrous

Ab out 150 g. of powdered anhydrous MgCl 250 ml of anhydrous MaCJ2. 2 and ether. were placed in a 500-ml . Morton flask equipped with a Stir-0-Vac high speed stirrer and reflux condenser . After the mixture had been

stirred vigorously for several hours , it thickened appreciably and could no longer be stirred effectively . Benzene was added . and distil

lation was carried out to remove ethyl ether. As the ether was being removed, the mixture became thinner and was more readily stirred. After

distillation of 400 ml . of benzene , the mixture was allowed to cool and

settle. VPC analysis showed only a trace of ether present in the super- · natant (£!. 0.01%) .

Forty milliliters of the mixture (as a slurry) was centrifuged,

the supernatant drawn off and 20 ml . of 1.3N . �-butyllithium in cycle hexane was added to the res idual MgCl2 . The mixture was shaken by hand and a heat of reaction noted. After shaking until the mixture became cool to the touch, Color Test IIa was positive . More of the slurry was added and the mixture stirred overnight. Color Test IIa was then nega

tive . Active alkyl analys is , total base, and a magnesium . analysis all

indicated that the supernatant solution was 0.6 M in di-sec•butylmagnesium.

f. From unsolvated �-product MgCl2. To 10 ml . of a slurry of

4. 8 g. (0 .05 moles) of MgCl2 (see following paragraph) in benzene con tained in a centrifuge tube fitted with a rubber septum was added 20 ml . of 1.3N �-butyllithium solution in cyclohexane . The mixture became warm on initial mixing of reagents, then cooled wh ile it was shaken 119

thoroughly by hand for 10-15 minutes . The mixture was centrifuged

and the supernatant , upon analysis , showed a negative Color Test IIa,

l.lON total base , 1.02N ac tive alkyl , 1.09N ma gnesium, and 0.03N chloride content . NMR showed the expected sextet of line s centered 43 5 cps . upfield from benzene indicative of the methine proton , and VPC analysis of a sample of the hydrolyzed product (organic layer) showe d n-butane as the only maj or hydrocarbon re sulting from hydrolysis . No alcoholic impur ities were found .

The MgCl2 used above was formed as a solid by-produc t of the reaction between n-amy l chloride and magne sium powder in benzene fol-

19 lowing the procedure of Glaze and Se lman . After withdrawal of the

supernatant di-�-amylmagne s ium solution in benzene , the MgCl2 residue wa s washed four times with benzene in the centrifuge tube pr ior to reaction with �-butyllithium.

g . From unsolvated ball-ffiilled lump commercial anhyd rous MgCl2 .

A we ight of 386 g. of lump anhydrous MgC l2 was partially ground with a mortar and pestle under nitrogen . It was then transferred to a 1-quart capacity jar containing smooth stones (1/2 to 1 inch diameter) ; approxi - mately 400 ml . of cyc lohexane was added and the jar was sealed , all under ni trogen. The mixture wa s rolled mechanically overnight on a jar mill. The finely ground portion of · the mixture was poured into lOO�ml . amb er bottles .under nitrogen!

To the contents of one of th ese bottles (100 ml ., 14 g. MgC l2) was adde d 85 ml . of 1.3N �-butyllithium in cyclohexane . The mix ture was stirred overnight . Color Test IIa was positive. The supernatant 120

solution on analy.s is , showed a 0 .llN magnesium and 0. 53N total base

content. Thus, the extent of exchange based on total alkalinity was

21 per cent.

h. · From alcohol�activated, desolvated commercial anhydrous

(�) Activation of commercial anhydrous MgClz ,!ill isopropyl atcohol . The remainder of the ball-milled cyclohexane slurry of

anhydrous MgCl2 ( 40 g.) described above was treated with 150-200 ml .

of isopropyl alcohol in a 500-ml . 3-necked round-bottomed flask equipped

with a Hershberg stirrer and jacketed Vigreux column with a variable

take-off fractionation head. (Lumps ,. 1/4 inch in diameter , of anhydrous

MgC1 2 may be successful ly substituted for the ball-milled slurry ; these

lumps readily break U:p to form a powder on trituration with isopropyl

alcohol.) Much heat was evolved and a considerable thickening of the

mixture occurred on addition of the alcohol . After about 1 hour of

vigorous stirring, 100 ml . of toluene was added and distillation was

begun to remove isopropyl alcohol as its toluene azeotrope. Successive

portions· of toluene were added to the residue to replace the solvent

distilled. After distilling about 450 ml . of toluene , a sample of

residual slurry was hydrolyzed. Analysis showed 1.1 moles MgC12 per

ml . of slurry and 5 mole. per eent. ieopropyl alcohol relative to MgC12 ·

Distillation to remove iaopropyl . alcohol. was continued; two 300-ml •

. portions of toluene were distilled. (Some reaction of MgC l2 was . . occurring since HCl appeared in the distillate.) · Analysis of the 121

slurry indicated it then contained 2.5 mo le per cent isopropyl alcohol

relative to the MgCl2 present .

('%)�· Reaction . of Ac tivated MgClz �· £-:butyllithium. To 175

. ili1. of the above slurry containing 6.5 grams (0.068 moles)• of MgC12

was. added 90 ml . of. 1. 24N �s- butyllithium in hexane . The reaction

'llaixture was stirred vigorously� first at 40°C for 1.5 hours, then at

room . temperature· overnight. Color Test. IIa was negative. The super-

natant solution: on analysis showed a 0.32N total base , 0.29N : active

, alkyl � and 0.28N Mg content.

i. From desolvated commercial MgCl "6H 0• A weight of 40.6 g. - 2 -2

(0 . 2 male) of MgCl2 • 6Hz0 was placed in a 3-necked 500-ml . flask equipped

with Hershberg stirrer and attached to a Dean-Stark trap • . Three hundred

milliliters of isoamyl alcohol was added and the stirred mixture was

heated to the boiling point . A total of 13 .5' ml . of water ,. from. a

theoretical amount of 21 .6 g. , was collected in the trap during a

period of 4-5 hours . No further removal of water occurred dur ing

several more hours of distillation. Isoamyl alcohol was then di stilled.

to leave a residual volume of 50 ml . The MgCl2 was soluble at the

distillation temperature , but upon cooling to room temperat:ure , the

. � entire mixture solidified. - One hundred milliliters of tetralin was

. . addedI to the solidified mixture and the contents o'f the flask . were

heated to reflux . The �ass again became a clear liquid. Tetralin

and isamy l alcohol were di.stilled. Another ioo ml . portion of

tetralin was added and the distillation was continued. A. crystal-

line precipitate formed du ring this stage of the distillation . The '. .,.

122

distillate was 'acidic indicating .some ·loss of· HCl, presumab ly by

· forming . an oxychloride· (ClMgOR) . After · hydrolysis of a sample · of ·

the slurry , -VPC of the, upper ·layer {on carbowax 400 ! at 150�) showed

·. only a trace of isoamy l alcohol to be present . Reac tion of th is

slurry with 100 ml . of a 1.20N solution of ��butyllithium in hexane

followed by analysis of the clear supernatant showed a negative

Color Test IIa and an 0.62N total base p.62N ·magnesium, and 0.44N

ac tive alky l content . The analysis indicated complete · exchange - of

Li and Mg took place, bu t tha t some of the Mg was present as alkoxide .

j . From commercial pmwdered anhydrous MgCl2 . Using powdered

anhydrous MgCl2 as obtained from Al fa Inorganics in place of the above

"activated" MgC12 , little or no ob servable exchange occurred with �-butyllithium in cyclohexane even with the high-speed stirring over

night.

2. Preparation of �-!,-_butylmagne sium

.... a. From dimethyl ether-solvated £l-product MgClz. Solid MgC l2

. (0 .08 mo le) , in one of the tubes from the experiment de scribed on page

was wa shed with benzene , resuspended in 40 ml . of benzene , and trans-

£erred to a 200-ml . 3-necked flask equipped with a magne tic stirring

bar . To the stirred slurry was added 83 ml . of a 1.44N !_-bu:tyllithium

solutio.n in pentane . The mixture was stirred for several hours and

then allowed to settle. Analyses on the supernatant solution showed a ,

negative Color Test IIa , 0.96N total base and 0.96N magnesiUm, indi-

eating complete exchange of Li and Mg. 123

b. From ether-desolvated MgC ll �-product from re action of �

Grignard reage nt !!!h benzyl chloride . To a slurry of 13.3 g. (0.14

mo les) of MgC l2 (see next paragraph) in 50 ml . of benzene was addedt.

with stirring , 50 ml . of 1.8N !-butyllith ium solution in benzene .

After the mixture was stirred for 45 minutes, Color Test I!a was

negative . Another 50 ml . of !-butyllith ium solution was added and

aga in Color Te st IIa was negative after 45 minutes of stirring. The

flask was then stoppered and placed in the refrigerator overnight.

NMR analysis of the superna tant solution showed only one !-butyl peak ,

no ethyl ether , and traces of pentane . For storage , about 150 ml . of

clear solution was withdrawn into an amb er bottle fitted with a serum

cap . Analysis showed 0.83N total base , 0.84N active alkyl and 0.82N magnesium content .

The MgCl 2 used above was prepared in the fo llowing manner . To

a Gr ignard reagent , prepared from 0.4 mo le of �-butyl chloride and 0.4

g. atom of magnesium (turnings) in 150 ml . of anhydrous ethyl ether , was added, at reflux and with vigorous stirring , 50 .6 g. (0 .5 moles)

of benzyl chloride (Baker analyzed reagent - 99 .8%) . A precipitate began to form immediately and copious gas evolution was no ted. (VPC

of the gas indicated th at it was 1- and 2-butenes and n-butane . ) After

addition was complete , the mixture was stirred thoroughly and then allowed to settle overnight . Gilman Color Test I on the supernatant

solution was negative; this solution was found by VPC to contain 12.5 mole per cent toluene , 70 per cent 1,2-diphenylethane , and 17 .5 per cent 2-methylbutylbenzene . The MgC12 residue in the reac tion flask 124 was wa shed five times with 100 �1 . portions of benzene , allowing the solids - to settle each time before removing the wash . The f.i fth wash wa s found to c.ontain about 5 w•dg}lt __per . .cent �thyf':�tbe·r".. ·The ·.slurry after the fifth wa sh on analysis showed a 5.4N Mg , 6.0N 01 , and O.OON

total base content . Continued washing of the solids did not remove all of the ethyl ether . The ether was removed by continuous codistil lation with benzene . The final ether content (VPC) of the supernatant

solution was found to be less than 0.02 weight per cent and the slurry was used as such for reaction with organolithium compounds in hydro carbon media.

3.. Preparation of !2.!,-�- �utylmagnesium

A weight of -25.5 g, (0 . 267 mole) of powdered anhydrous MgCl2 and

100 ml . of anhydrous ethyl ether were placed in a 200-ml . round-bottom,

3�necked flask equipped with a magnetic stirrer and reflux condenser .

To the stirred, cooled (ice bath) mixture was added all at once 30 ml . of 3.7N-n-butyllithium in benzene . After two hours , Color Test IIA was negative . The mixture was stirred and allowed to come to room tem perature overnight . Aliquots of the clear supernatant solution were withdrawn for analysis and found to contain 0. 76N 'total base, 0.76N active alkyl and 0.88N chloride . Thus , the product appeared to be in the form of n-butylmagnesium chloride ,- not di�!-butylmagnesium . Approxi mately 80 ml . of the clear supernatant solution and 50 . ml . ;;_of cyclohexane were transferred to a dry 200 ml ., 3-necked, round�bottom flask. flushed with. nitrogen and fitted with a rubber septum and a vacuum-:- jack'eted 12 125 inch Vigreux column with variable take-off fractionation head and receiver . After distillation of 50 ml . of solvent and additon of anether 50 ml . portion ef cyclohexane , the head temperature reached

78°C . This process of distillation and replenishment of · solvent to . the residual solution was continued until a total ef 250 ml . of sol• .. vent had been distilled. The residual solution was allowed to coal to room temperature . A total of 71 ml . of a quite mob ile supernatant solution was withdrawn (some white solid residue was left) and found .. to contain 0.. 44N total base, 0.42N. ..ac tive alkyl and O.OlN Cl . VPC analysis of a hydrolyzed samp le of the solution showed that it. contained less than · 2 mole per cent ethyl ether · relative to the di-n- butylmagnesium . Thus , about 50 per cent of the original active alkyl content remained in solution after removal of the ethyl ether.

4. Preparation .o.f . Di ..neop .entylmagnesium ___.�______-- ...... � - - To 147 ml . of a solution of 0.47N neopentyllithium (0.069 mo les) in a pentane-dimethyl ether mixture (£!. 10: 1) contained in a 3-necked,

200-ml . round-bottom flask equipped with magnetic . stirring bar and reflux condenser , and cooled in an ice bath , there was added 13 g.

(0. 137 mo les) of powdered anhydrous MgC l2 . After the mixture had been stirred for several hours , Color Test IIa was negative. Some evaporation of solvent occurred during reaction. Analysis of the supernatant solutien showed an 0.55N active alkyl, 0.56N total base and 0.54N Mg content. 126

D. DIRECT PREPARATION OF DIALKYLMAGNESIUM REAGENTS

FROM MAGNESIUM METAL AND ALKYL CHLORIDES

1. Preparation � ��- butylmagnes ium

a. � dimethyl ether-cyc lohexane solution. A weight of 2.43 g. (0.1 g. atom) of magnesium turnings and a few crystals of iodine were placed in a 500-ml ., 3-necked flask fitted with a thermometer

(-100 to +50°), mechanical stirrer, Dry Ice condenser and graduated dropping funnel. Enough' DME (dried by pas sage over silica gel) was condensed into the flask to cover the turnings and 1-2 ml . of �-butyl chloride , from a total of 9.3 g. (0.1 mo le) was added. After addition of a few drops of ethylene dibromide, the reaction appeared to begin as indicated by the disappearance of the iodine color and the develop- ment of a cloudy solution. The remainder of 100 ml . of DME was con densed into the flask. at -25°C and the remainder of the .�-butyl chloride was added. Reaction was slow and 100 ml . of cyclohexane was added gradually to the flask. The mix ture was allowed to warm slowly to room temperature and allowed to stir at room temperature for two days .

A wh ite slurry formed during this time . The reaction mixture was heated 0 to 45 C for 2 hours , result ing in almost complete disappearance of the magnesium turnings. The reaction mixture was allowed to cool and settle overnight . Analysis of the clear supernatant solution showed a 0.77N total base and 0.085N chloride content . Thus, th& yield of di-�-butylmagne s ium appeared to ·be quant itative; the NMR ' spectrum of the solution wa.s identical with a similar solution prepared with the authentic magnes ium alkyl. 127

b. � vacuum- stripping � solvent 1!2m !h! Grignard reagent ,

followed ��-solution in benzene • . A solution of Grignard reagent,

prepared from 9.7 g. (0.4 g. atom) of magnesium turnings and 37 g.

(0.4 mo les) of �-butyl chloride in 150 ml . of anhydrous ethyl ether ,

showed upon analysis a total base concentration of 2.20N and an ac tive

alkyl concentration of 2.13N. Eighty-five milliliters of this solution was trans ferred to a 200-ml ., 3-necked round-bottom flask equipped with

a capillary tube for admitting nitrogen below the liquid level and a

distillation head leading to a receiver and vacuum line. The flask was heated wi th an oil bath . Solvent was removed under vacuum to about

3-4 mm pressure with a bath temperature of 75°C. The mixture eventually became qu ite viscous and slowly evolved ether. The residue was evacuated for 12 hours at this temperature . After returning to room temperature and atmospheric pressure.,, the flask contents , a viscous mass, were stirred with about 50 ml . of benzene. The mixture thinned out, with a granular wh ite precipitate being evident in the mixture . On analysis, an ethyl ether to di-�-butylmagnesium mo lar ratio of 0.7 was found. The solution had a total base and ac tive alkyl concentration of

1. 7N and negligible Cl content . It was desolvated by continuous codistillation with a hydrocarbon solvent as previously described.

c. In diethyl ether-benzene solution . A weight of 10 g. (0. 415 g. atom) of magnesium turnings were activated with iodine and mixed with

200 ml . of benzene and 50 �1. of anhydrous et��l ether , The mixture was ·. heated to reflux and 42 ml . (0.39 moles) of _!-bu-tyl chlor ide added por- . tionwise to the vigorously stirred mixture . After 16 ml . qf the halide 128

. had been added, reaction was init iated by addition of a sma ll portion

of ,!-butyl magnes ium chloride in ether . The reaction proceeded vigor

ously and intermittent cooling with an ice-bath was necessary . When

the initial reaction had subsided, slow ·halide addition was cont inued.

The re.action mixture was heated at reflux (770C) for 1 hour and then

allowed to cool to room temperature . No MgCI 2 precipitate was in evi

dence at this time . Total base concentrat ion was found to be 1.33N

indicating the yield of ,!-butylmagnes ium chloride to be essentially quantitative . The solution was allowed to stand overnight , still no

precipitation was evident .

Thirty mil liliters of this solution was placed in a distillation apparatus as described above and desolvat'ed . as previously de scri.bed.

Slow precipitation of MgCl2 began as soon as ,solvent started distil ling . When between 10 and 15 ml . of solvent had been removed, the mixture was allowed to cool and settle. Analysis of the supernatant showed a 2.48N total base and 1.85N chloride content . The .molar ratio of ether to carbon bound magnesium was 2: 1. Thus , removal of 40-50 per cent of the solvent resulted in precipitation of 25 per cent of the MgClz, leaving an equivalent amount of di -!!£-butylmagnesium in solution.

2. Preparation of Di-�.�.tylmagne sium �· weight of 2 . �3 g •. (0.1, g . at.Oln�, Qft-lll4&D,e..eium ·��al · powder

(40 me sh) was ·he�d llith a cry�al of iodine. in a few ml . o� �yclo l\e.¥ane . ·A few nil . of a so�ution .of 9.?5 g. (0.1 U1ole) of· n,-but;yl chloride in 80 ml . of cyclohexane was added to the mixture . The iodine color disappeared and a cloudiness formed in the solution . The mixture was heated to reflux and the rema inder of the halide solution was added over a per�od of 2 hours wi th vigoraus stirring with high speed stirrer . Stirring and refluxing were continued for several hours after additon of the halide solution was complete. A white crystalline

,• precipitate formed on the walls of the flask du ring .this time . After the reaction mixture was cooled, the whi te precipitate was found to contain active alkyl , but none was found dissolved in the supernatant solution . DME was slawly bubbled into the mixture until all the �hite solid had dissolved. The mixture was allowed to settle and the clear yellow supernatant solution found to contain negligible Cl and to . be

0.58N total base and in ac tive alkyl . ·'.The yteld:rof dba-butylinagne.s ium was thus £!· 60 per cent .

3. Preparation 2f B!-�·aeylmagnes ium

Into a 200-ml ., 3-necked, round-bottom flask equipped with Hershberg stirrer , reflux condenser , and 100-ml . graduated dropping funnel was placed

2.68 g. (1 .1 g. atom) of magnesium turnings (under nitrogen) and a cryst'al of iodine . The flask was heated to 85° by means of an oil bath and 1-2 ml . from a total of 10 .6 g. (0.1 mole) of 1-chloropentane was added.

The reaction began readily with disappearance of iodine color and the appearan�e of. fuming . The remainder of the halide· was added over a 1 hour period wh ile vigorous stirring and heating at reflux was maintained:

About 1 hour after addition was complete, everything in the flask became _ dry and lumpy . Fifty milliliters of benzene was added with ·vigorous

·. 130" · stirring and heating at reflux . After one hour another 50 ml . of benzene wa s adde d and stirring and heat ing continued for an addi• tional 3-4 hours � The mixture had become viscous during th is time . '

.; and was allowe d to cool and settle overnight . 'tlte mixtuz:e �as centrifuged. On analysis , the clear,colorless, viscous supernatant solution showed an 0.27N total base and 0.25N ac tive alkyl content , but no re s idual 1-chloropentane . n-Pentane was the only new organic product found after hydrolysis .

4. Prepara tion of Diisopr opylmagnes ium

.A weight of 9.7 g. (0 .4 g. atom) of magne sium tu rnings were activated wi th iodine and covered with 100 ml . of dry benzene . Ten grams (0.2 mo le) of dime thyl ether was condensed into the mixture and 2.3 ml . of isopropyl chloride was added . No react ion occurred .

After one -third of a solution of the remainder of 31.4 g. (0 .4 mo le) of isopropyl chloride in 125 ml . of benzene had been added, reaction began as evi denced by a rise in temperature . Intermi ttent cooling with an ice bath kept the temperature between 35 and 45°C wh ile addi• tion of the hal ide solution wa s continued over a one hour period.

During this time an addi tional 34 g. (0.74 mo les) of dime thyl ether was also added. The react ion mixture was stirred and heated to 50° for two hours . The mixture was then allowed to cool and settle overnight . One hundred fifty mi lliliters of the clear colorless supernatant solution was removed and the rema inder of the slurry of MgCl 2 and unreacted magnes ium turnings wa s tr�nsferred to two 50 ml . centrifuge tubes and centr ifuged. Sixty milliliters of the supernatant 131 solution was removed and comb ined with the 150 ml . of product solution.

Th is solution showed 1.45N total base and 1.35N active alkyl , for an apparent yield of 80 per cent of diisoprdpylmagnesium.

E •.PRE PARATION OF MIXED SOLUBLE DI.,.n.-BUTYLMAGNESIUM

COMPLEXES WITH _!-BUTYLLITHIUM AND !!,-BUTYLLITHIUM

1. Preparation .Qf Di-n-butylmagneaium ..

A weight of 16.7 g. (0 . 176 mole) of powdered anhydrous MgCl2 was mixed with 75 ml . of anhydrous benzene in a nitrogen-flushed 200-ml .,

- 3-necked, round-bottom flask equipped with reflux condenser , serum cap, and magnetic stirring bar. Thirty milliliters of anhydrous ethyl ether

(0 .3 mo les) was added followed by 30 ml . of 3.7N !!,-butyllithium in ben� zene·. The mixture became warm and was cooled and stirred in an ice bath for 1-2 hours. After allowing the mixture to come to room temperature, it was stirred overnight; analysis of the clear supernatant showed a negative Color Test IIa and 0.8N total base and 0.3N Cl content . One hundred ten milliliters of the clear supernatant solution was trans- f'erred to :a 200-ml . flask fitted for di stillation as described above and 9.2 ml . of 3.7N !!,-butyllithium in benzene added to react with the chloride ion present in solution. After it was stirred 30 minutes , the solution was analyzed for chloride ion. None was found. The ether was distilled using benzene as codistillate; solid di-!!,-butylmagne&tum precipitated during the distillation. After 200 ml . of benzene had been distilled, ether was still found (VPC) to be present in the slurry . The slurry was then heated to 120° , almost to dryness, while 132

another 70-80 ml . of distillate was collected. Benzene was added to

bring the total volume to 100 ml . wh ile the mixture was allowed to

stir and cool to room temperature.

2. Preparation of Butyl lithium-Butylmagnesium Complexes

Three 5 ml . aliquots of the stirred slurry were placed .in

nitrogen-flushed centrifuge tubes fitted with serum .caps . The follow

ing reagents were added to each tube : · 0.81 ml . of 1. 24N. !,-butyllithium

in cyclohexane (tube No •. 1) ; 0.625 ml . of 1.6N �-butyllithium in hexane

plus 0.2 ml . of cyclohexane (tube No . 2) ; 0.81 ml . of cyclohexane (tube

No. 3) . The tubes were shaken we ll , centrifuged and the supernatant analyzed for total · base by titration and for ethyl ether content by

NMR . The total base contents in each of the three tubes were as follows :

4.70 meq. in tube 1; . 4.47 meq. in tube 2 ; 0.70 meq . in tube 3. The res idue in tube 3-was di ssolved in THF and found to contain 3.88 meq .

total base . Thus the total base content present in 5 ml . of original slurry was 4.58 meq. The percentage solubilizations of the di��-butyl magnesium by .�- and !,-butyllithium reagents calculated from these data were 76 and 81 per cent respectively . Subtracting the amount of di-� butylmagnesium itself soluble in this medium {determined with tube 3 above), the mole ratios of di-�-butylmagnes ium to �- and !,-butyllithium reagents in the solutions were found to be 1. 9 and 2.0 respectively . 133

F . PREPARATION OF SOLUBLE , MIXED DIALKYLMAGNESIUM COMPLEXES

FROM INS OLUBLE DIALKYLMAGNESIUM COMPOUNDS

1. Preparation .Q!.!.Sol uble Diisobutyl ·lli:,-,1,- bYtylmagnesium,Comp lex ·

To 0.35 g. (5.0 meq .) of solid diisobutylmagnesium (prepared by reaction of isobutylmagnesium chloride with isobutyllithium in ethyl ether followed by desolvation with cyclohexane) was added 5 ml . of a

1.35N di-!,-butylmagnes ium solution in n-hexane and 2.5 ml . of cyclo hexane . Complete dissolution of the diisobutylmagnesium (which was previous ly found to be soluble to the extent of only 0.09 moles/liter in cyclohexane) occurred yielding a solution having a total base concentration of 1.56N, which may be compared to a calculated value of 1.57N.

2. Preparation .Q!.!.Solu ble Diisobutyl ·£i-�-butylmagnesium Compl ex

To 0.8167 g. (10.86 meq;) of sol id diisobutylmagnesium sus pended in a few ml . of benzene was added 7.54 .m l. of 1.57N di-!,-butyl magnesium in benzene (desolvated) . After thorough mixing and centri fugation; the clear supernatant solut ion was analyzed for the ratio of

!-butyl to isobutyl groups by NMR . An approximate 1:1 ratio was found, based on the ra·tio of !-buty l protons to me thyl protons of the isobutyl group.

3. Preparation .Q!.!.Solu ble Bi�butyl -B!·�- butylmagnesium Compl ex

To a slurry of 13.3 g. (0 .14 mo les) of activated MgCl2 (ether free) in 50 ml . of benzene was added with stirring 40 ml . of 1.75N

!-butyllithium in benzene . After stirring the mixture for 15-20 minutes, 134

19 ml . of 3.7N -.n:·butyllithium in benzene was added . The mixture imme· ·

' diately became quite warm. After stirring until cool , analysis of the

supernata�t solution showed a negative Color Test IIa and 1.03N total base,l .02 active alky l and 1.04 Mg contents. NMR analysis showed a

!·butyl to-n-butyl group ratio of l:O.S.

G. INSOLUBLE COMPLEXES OF DIALKYLMAGNESIUM REAGENTS WITH

TRIETHYLENEDIA."f!NE AND METHYLTRIETHYLENEDIAMINE

1. Comp lexes �_ Triethvl enediamine

a. Di�!a�butylmagne sium�triethylenediamine complex. To 5.0 ml .

of a O.SSN solutian a£ di-sec-butylmagnesium (2 .13 lomo1. al) in benzene· cyclohexane (desolvated) in a nitrogen-flushed dry centrifuge tube

fitted with a serum cap was added 3.0 ml. of an 0.69M- solution of tri-

ethylenediamine (TED) in cyclahexane (2.07 11DC)le�'�· A crystalline precipitate formed immediately. After shaking the mixture thoroughly ,

it was centrifuged and the solids washed twice with cy.clohexane . NMR analysis of a solution of the solid in THF showe d • triethylenediamine

(all protons equivalent) to di-�-butylmagnesium (me thine proton) ratia of 0.9 ± 0.·1 .

In another experiment, 3.2 ml . of 0.69M TED (2 .2 mmolee) in cyclohe�ane was added to �.10ml . of 0.98N di-�-butylmagnesium (4 .4 mmoles) in benzene . A fine·grained white crystalline precipitate was

formed wh ich was centrifuged and the solids were washed three time s with .u-pentane . The centrifuge tube was heated in an oil bath at 135

80-90° and the solids blown dry with nitrogen. The tube was trans-

ferred to a glove bag and some of the dry solid transferred to a dry

nitrogen-flushed me lting point capillary. The melting point capillary

was quickly sealed in a flame . The solid me lted with decomposition

0 (gas evolution) at 201-203 C. The remaining solid in the centrifuge

tube was weighed on an analytical balance (0.5159 grams) and dissolved

in 5 ml . of anhydrous THF . Ac tive alkyl analysis of this solution

showed the presence of 2.09 mmo les of di-�-butylmagnesi� . The equiv

alent we ight of the solid thus obtained was 248 (theory for (C4H9) 2-

Mg · C6H12N2 , 250) .

b. Other dialkylmagnesium- triethylenediamine complexes . I� a

similar manner there were prepared 1: 1 complexes of triethylene.diamine

and di��-butylmagnesium and di-!-butylmagnesium. No crystalline precipitate was formed when di-�-amylmagnesium solution in benzene was

treated with an excess of 0.69M TED solution, regardless of cooling or concentration of the solution. Diisobutylmagnesium and diisopropyl- magnesium �ere too insoluble in hydrocarbon solvents to permi� the formation of crystalline precipitates with TED. · On the other hand, in the presence of enough dimethyl ether to permit their solubiliziation in hydrocarbon solvents, they formed non-stoichiometric crystall ine complexes with TED containing dimethyl ether. Attempts to obtain TED complexes of a mixed magnesium alkyl from a benzetle+>Cydlahexane ··s1il.'U- tion of !•butyl��-butylmagpesium resulted in preferred complexation of di-!-butylmagnesium accompanied by a co-precipitation of insoluble di-

�-butylmagnesium . 136 2. · Compl exes !iSh Methyl triethylenediamine

a. ]1-&·hutylmagnesium-methyl triethyl enediamine compl ex. A

volume of 5.5 ml . of 0.85N di-�-butylmagnesium solution in benzene-

cyclohexane {desolvated) was added to a centrifuge tube containing

0.3 ml . of methyltriethylenediamine (MTED) dissolved in 3 ml ; of

benzene (7.• 92 mmoles/ml .). The fine crystalline precipitate wh ich

formed slowly was centrifuged and the supernatant solution drawn

off. The solid {0.54 g.) was recrystallized from 8-10 ml . of benzene

to give monoclinic prisms , which on washing and drying, (as de scribed

for TED complexes) me lted at 160°C (dec .) in a sealed tube. NMR

analysis of a solution of these crystals in THF showed the presence

of a 1: 1 (!-Bu) 2Mg ·MTED) complex based on integration of the methylene

protons of the MTED and the me thine proton of the di-�-butylmagnesium .

b. Other dialkylmagnesium-meI thyl triethylenediamine compl exes.

SiuvUarly there was prepared the 1:.1 di-1£!!.-butylmagnesium-MTED com-

plex . Recrystallization from benzene gave star-shaped prisms , m.p.

. \ 163° (dec.) . No crystalline pre� ipitate was formed between MTED and

di-�-butylmagnesium in benzene-cyclohexane regardless of cooling or

concentration of the solution. MTED complexes of mixed magnesium ·

alkyls such as �-butyl-isobutylmagnesium and methyl-isobutylmagnesium yielded non-stoichiometric complexes which melted over a wide range;·

NMR analysis showe d a molar ratio of �-butyl to isobutyl groups in

the · former complex to be 1.62, wh ile the molar ratio of methyl to

isobutyl groups in the latter complex was found to be 1.48. In the 137 case of the methyl-isobutylmagnesium TED complex, the average ratio of dialkylmagnesium reagent to MTED was 4: 1, indicating -a breakup of the mixed .a l kylmagnesium with preferred complexation of dimethylmagnesium and coprecipitation of uncomplexed diisobutylmagnesium .

H .• METALLATIONS EFFECTED WITH DIALKYLMAGNESIUM REAGENTS

1. Me==tallat=--=;;.:.;;;;ion;..;;:,;;.;. -of Re sorc..in ol Dimethyl ·-Ether

a. In the absence of base . Thirty-four milliliters of 0.77N di-�-butylmagnesium (13.5 mmo les) in hexane-cyclohexane was placed in a 50-ml ., 3-necked, conical-shaped flask which had been flushed with nitrogen and equipped wi th a reflux condenser, magnetic stirring bar and rubber septum on one neck. A volume of 3.8 ml . (27 mmo les) of resorcinol dimethyl ether (DMR) was added to the flask, stirring begun and the contents heated to reflux for 16 hours. Large lacy crystals formed in the mixture. After allowing the mixture to cool , the wet crystals were transferred under: .nitrogen to a centrifuge· tube

(using a Glove Bag to effect the transfer) and dissolved in 1 ml . of

THF . The clear THF solution formed was analyzed by NMR; the aromatic region showed two signals, a 2: 1 ratio of protons ortho and � to methoxy groups in metal lated resorcinol dimethyl ether ; the aliphatic region showed a methine proton (like that in di -s-butylmagnesium) in a 1:1 ratio to the � proton� NMR, analysis of the supernatant solution from these crystals showed only unreac ted resorcinol dimethyl ether

(IMR) to be present, no df-�-butylmagnesium . (The remainder of the

.• 138 . reaction mixture was carbonated by pouring it into a Dry Ic e-ether slurry , but evaporation of the carbonated slurry on a steam bath was ' inadvertantly carried to dryness, wh ich caused thermal decomposition of the salt to DMR and MgC03 .)

b. In .the pr esence of base.

(1)· ![[·and THEDA . Employing the same procedure as described above, on a comparable scale, separate runs were carried out in Which

2 mole equivalents of THF and in wh ich 2 mo le equivalents of TMEDA per mo le of magnes ium reagent were each added to the reaction solutions .

No metallation occurred.- NMR analysis showed only unre'ac ted starting materials to be present after 16 hours reflux of the react ion solutions .

(2) THF plus po tassium-�-butoxide. To the above run containing

THF , after the 16 hours of reflux , there was added an amount of solid potassium !-butoxide equal , in mo les , to twice the moles of di -�-butyl-

' magnesium present . The mixture immediately turned red and then, after a few minutes almost -black. A sticky precipitate formed. The reaction mixture was carbonated by pouring it into a stirred Dry Ice-ether slurry . The ether was carefully evaporated on a steam bath and the res idual dry solid was washed three times with ether in a centrifuge tube fitted with a rubber septum . Some of the slurry was removed and the solid remaining in the tube was washed twice more with ether, and then blown dry under nitrogen. The solid was dissolved in D20 and after centrifugation, the clear supernatant solution was examined by

NMR. The salt of 2,6- dimethoxybenzoic acid was found to be present

t as well as the salt of 2-methylbutyric acid, preslllli�bly as ' heir 139 potassium salts ; integration of the aroma ti c protons in the 2,6- dime thoxybenzoic ac id salt and the methine proton in the 2-methylbutyric acid salt gave a ratio of the two salts (aromatic: alipbatic) of 57:43.

(3) TMEDA £1us li.t.ll tUUI, 1�.b�.t.oii9..e . To the run in . (l)· above containing TMEDA , after the initial 16 hours of reflux, there was added a molar excess of lithium !-butoxide (relative to the moles of

R2Mg present) as a hexane slurry . Only a slight deepening in color was noted. The solution was heated to reflux for 16 hours . During this time gases were evolved and a wh ite precipitate formed. On cooling , a sample of the slurry was centrifuged. NMR analysis of the clear supernatant showed the presence of met•lla�ed preduct �(S!. l.a. above) as well as unreacted DMR. The solid residue , although sparingly soluble in THF , also showed the presence of metallated DMR

/ by NMR analysis. The reaction mixture was carbonated by pouring it into a stirred Dry Ice-ether slurry and then the carbonated mixture was ·allowed to evaporate to dryness at room temperature. A portion of the residue was washed twice with ether and mixed with 3 ml . of

D20. After centrifugation , an NMR exa�ination of the supernatant

solution showed the presence of the sal. ..\ (presumably the lithium salt) of 2,6-dimethoxybenzoic acid, but no 2-methylbutyric acid salt.

The yield thus appeared essentially quantitative.

z.. Metal lat. ion.-� Thioph ene , Me;t�l��d ·. a�o· . t!7n .without Lewis �· A solution of 25 ml . of 0. 77N di-�-bUtylmagnesium and 1.5 ml . (1 .6 g. , 19.3 mmoles) of thiophene in 140

hexane-cyclohexane was stirred overnight. NMR· analysis of the solution

showed no me tallated thiophene and no loss of the magnesium alkyl . Heat

ing the solution overnight at reflux resulted in the formation of a

sticky precipitate on the walls. NMR analysis of both the supernatant

solution and a solution of the precipitate in THF showed the 'absence of

di-!,-butylmagnesium. The THF solution contained metallated thiophene .

Carbonation and work-up of the solution in the usual manner produced

2-thiophenecarboxylic ac id in at least 34 per cent yield; a portion

was inadvertantly lost during work up .

b. At tempted metallati.on �.Lewis �· Addition of a molar

excess of tetrahydrofuran (relative to the di-!,-butylmagnesium) to

another reaction mixture of the reagents used above did not effect the

/ me tallation of the thi0phene· after stirring for 16 hi!UTS' at raom,tem

perature nor after 12 hours at reflux. THF solutions of the 1: 1 com

plexes of di -_!-butylmagnesium and of di-!,·butylmagnes ium with triethylene

diamine were also ineffectual in metallating thiophene even with heating

at reflux for 24 hours.

c. Me tallation � THF plus po tassium �-butoxide . To the

unreacted mixture in b. above containing thiophene , di-_!-butylmagnesium

and an excess of THF , there was added in solid form, 2 moles of potassium

!,-butoxide per mo le. of di-!_-butylmagnesium . The mixture innnediately

turned red, then almost black , and generated heat . A considerable amount

of black ma terial precipitated. The mixture was carbonated and worked

up to obtain a D20 solution of carboxylic ac id salts in the usual manner . 141

NMR examination of the D20 solution of salts showed the presence of

2-th iophene carboxylic ac id (presumab ly as the potassium salt) and

: ,t· butyl alcohol .. Only a very small amount of 2•methylb\ityric acid

salt was found . The ratio of aromatic to aliphatic ac id was about

10: 1.

d •• Metallatioq.!!1 ill. presence .2! &-butyl lithium · plus .TMEDA. To 2 ml . of a solution containing 0.8 mmoles of di-a-buty lmagnes ium

. and 0.4 mmoles of. �·butyllithium in a mixed hydrocarbon solvent there · was added O.l7 ml . ( 2 mmoles) of thiophene and 0.2 ml . (1 .3 mmoles) of TMEDA . After standing at room temperature for 5 Q&ys , the supernatant solution was examined by NMR. Metal latedthiophene was present

as we ll as . dt�·butylmagnesium. NMR signals of the latter could be

seen clearly as an upfield triplet in the "organmnetallic region"

around TMS. No sextet of lines indicating a �-butyl group could be

found in that region.

3 . ff@tal lat ipn; .21 .l·Picoline

In an experiment comparable to l.a. above, 1.80 g. (1 .90 ml .,

19 2 .3 mmoles)· of Y-picoline was a•ded to. 5 ml .· of a 0.77N di-s--butylmagnesium in hexane-cyclohexane solution at room temperature . An immediate red color developed and deepened over a 16 hour period. A gloasy dark solid had formed during this time . .The supernatant solution was examined by NMR and found to contain a 3: 1 mb:ture of metallate'd to anmet•llated Y -picoline , as determined by integration of the methylene and �ethyl protons respectively . (The methy lene protons were shielded 142

with respect to methyl protons by 25 cps) . The methine protQn sex�et

of the �-butyl gropp in the original magnesium reagent was ab sent frem

. the spectrum . The gloasy solid was dissolved in .THF but a sticky mate

rial formed and coating the wal ls of the NMR tube , precluding takins a

spectrum .

Both the original supernatant so lut ion and the THF · solution were

. added to a pre-cooled (0°C) solution of 3 .0 ml . (3 . 6 g� , 26 DDDoles)• of

benzoyl chloride in 15 ml . of ether . A precipitate formed and the mix

ture became quite thick. More THF (20 ml .) and benzoyl chloride· (5 ml. )

were added and the mixture became less viscous . After stirring for 0.5

hr ., water was added to hydrolyze the mixture . The layers were separated

and to the water solution there was carefully added saturated aqueous

Na2C03 solution unt il gas evolut ion seemed to cease. The water layer

was then extracted with ether . This ether solution still contained

some unreacted benzoyl chloride ; it was evaporated on a steam bath

leaving a dark bro� oily residue wh ich did no t crystallize on cooling .

Th is residue was extracted with hot �-hexane and on cooling , white .

crystals formed along with some dark viscous oil . The crystals were

dried and found to ·melt at 113-117° (lit. m. p. of y-phenacy lpyridine ,

115°C) . The remainder of the crystallized residue was dissolved in

THF . NMR examination of th is solution showed that the dissolved mate

rial was Y-phenacylpyridine .

4. M!.t alb t!PP. .2£;. Toluene

a� . At.t.ems>t.e.d. jn,e.t!l lati.e�n. A vo lume of 100 ml . of 0. 77 N di-�

butylmagnesium in hexane-cyclohexane was distilled under vacuum to 0.02 143 ° mm . and 40 C to give a light yellow oily liquid. After returning the

sy stem to atmospheric pressure with dry nitrogen , 50 ml . of toluene was added and the mixture stirred thoroughly. Analys is for total base

showed that the toluene solution was 1.33N . This di-�-butylmagnes ium had never been in contact with base, i.e. , it was not produced by

desolvation.

Twenty milliliters of the toluene solution· and 4 ml . (26 .6 mmoles) of TMEDA were mixed together and heated to 75°C for 16 hours .

No change in appearance of the solution was noted. The solution was

then heated to reflux for 16 hours . NMR examination showed no loss of di-�-butylmagnes ium relative to TMEDA. The solution was carbonated and the sal ts were evaporated to a dark red viscous residue on a steam bath . The residue was then heated in an oil bath at 150° for a short

time ; on cooling it solidified to a red glass . To a portion of this glass was added D20 and Na2co3 , but very little solution appeared to occur . Benzene and hexane were added; the mixture was stirred well and then heated in an oil bath to remove the organic solvents . The resulting aqueous (D20) suspens ion wa s centrifuged and an NMR spec

trum taken on the supernatant solution. Only the salt of 2 -methyl butyric acid was found; there �re no NMR signal s in the aromatic region .

. b. Me tallation � pa rtially pyrolyzed di-&-butylmagne sium.

Volume s of 27 ml . of toluene and 15 ml . of a 1.68 M solution of di-� butylmagnesium (25 mmoles) in u-hexane wh ich had been partial ly de sol vated of DME ,were added to a 50-ml . nitrogen-filled flask equipped for distillation with a Vigreux column . Ten mil liliters of hexane was 144

distilled slowly, during wh ich cracking of di-�-butylmagnesium was

apparent ; a white precipitate formed and fume s were evolved. The mix··

ture was allowed to cool and the residual solution was analyzed by VPC

for the presence of hexane. Ab out 5 wt . per cent hexane was still

present in the solution. A volume of 7.5 ml . (50.1 mmoles) of TMEDA was added to the flask and the mixture was heated to reflux. After

2 hrs . the mixture was allowed to cool and settle. Analysis of the

supernatant solution wh ich was a pinkish-red in color, indicated that

about 75 per cent of the di-�-butylmagnesium ·had reacted; no TMEDA

could be found in the supernatant . The reaction mixture was carbonated

in 200 ml . of a slurry of ether and Dry Ice . The ether slurry was treated with excess 6N HCl and the organic layer separated. The aqueous layer was washed twice with ether and the washings combined with the main layer .

The ethereal layer was extrac ted with 30 ml . of 10 per cent aqueous NaOH

and washed twice with water . The combined extrac ts were filtered, ac idi

fied with 6N HCl and extracted three times with ether. The ethereal extrac ts were dr ied over Drierite and evaporated on a steam bath . The res idue , which partially solidified on standing , weighed 2.45 grams and was analyzed by NMR. The ratio of phenylacetic to a-methylbutyric acids

in the residue was 1.6 and represented a yield of phenylacetic acid of

22 per cent , based on the original �-butylmagnesium content .

5. Me tallation. of Benzene � Partially Pyrolyzed di-&·Butylmagnesium

To 13 ml . of a 0.73N di-�-butylmagnesium (4.8 mmoles) solution .in benzene was added 1.4 ml . (9.5 mmo les) of TMEDA . During the original de solvation of the benzene solution, pyrolysis had been noted. The 145

mixture was refluxed. Gas evolution, as well as the formation of a

.yellew,.. sticky precipitate, was noted. After 4 hrs. , the mixture was

coeled; analysis showed that about one·half of the di.-s-·butylmagnesium I had reacted. The reaction mixture was carbonated in a Dry Ice-ether

slurry and worked up as in 1 above . NMR.analy sis of the residue

showed the presence of ab out 14 per cent benzoic acid and 86 per cent

a-methylbutyric acid; the yield of benzoic acid, based on initial

A·butylmagnesium present, was about 10 per cent.

6. Metallation 2! Alkxnes

a. Acetylene . The acetylene generated from.the hydrolysi� of

4 grams of lump CaC2 (one gram at a· time) was passed, with the aid of

a nitrogen gas purge , through a tube of barium oxide and into 20 ml .

of benzene contained in a 50 ml . 3-necked flask equipped �ith magnetic

stirring bar , reflux condenser, gas inlet tube, and rubber. septum . To

the flask there was added 26 ml . of 0.77N di-A-butyl�gnesium solution

in hexane-cyclohexane . A light green, milky mi,ture resulted. Three more 1 g. lumps of CaC2 were hydrolyzed and the gas so generated swept

into the above mixture. The precipitate became more particulate and

the mixture turned a deep lavender or purple. After it had been stirred

for 30 min. , the mixture was heated to boiling for 30 min . and allowed

. to cool and settle . The solid retained the rp�l·e color . D20 was

added- to the contents of the flask and the gases evolved collected in

.a vacuum system and chromatographed on a special triethylamine col�

0 at -78 • No ,!!•·butane was found, and the acetylenes obtained were : 146 acetylene , 7 per cent , monodeuteroacetylene , 28 per cent, and

dideuteroacetylene , 65 per cent .

b' . 1-Methxl-!- butvne •. To 35 ml . of 0. 77N di-,!.- butylmasnesium

(13.5 mmo les) in hexane-cyclohexane was added 1.3 ml . (0 .92 g. ,_ 13.5 mmoles)• of 3-methyl-l�butyne . Heat was generated, but no precipitate

formed nor was any gas evolved. After it cooled the solution was

analyzed by NMR . A shift of about 10 cps of the ·methine sextet (of

the. a-butyl group) was noted, but the rest of the spectrum was masked - ° by solvent. The solvent was removed under vacuum (0.02 mm and 40 C) ,

and the residual viscous yellow oil was di ssolved by addition of 10 ml . of benzene. An NMR spectrum of the benzene solution showed loss

of the. !.-butyl methine sextet in the expected region (9-9.5t) and gave

a complicated pattern cons isting mainly of eighteen lines between 6.2

c. !-Methyl -!-pentyne. To 3 ml . of 0.63N di�t-butylmagnesium

(0.95 mmo les) in benzene contained in a centrifuge tube. fitted wi th a rubber -septum was added a solution of 80 mg . (0. 98 mmoles) of 4-methyl•

1-pentyne in 2 ml . of benzene . After a few minutes the solution had become warm , but no gas evolution occurred; there was a slight hazi- ness in the solution. After centrifugation the clear solution was examined by NMR . Two main sets of doub lets appeared between 7.9t.r and

8.6t, and _three much smaller broader 'peaks appeared at 7.2, 7.5 and 7.8t.

The smaller lower field doublet had an area of about 1/2 that of the upper field doublet. The solution was carbonated in a Dry Ice-ether 147 slurry and on workup , the solution of salts in D20·was examined by

NMR . A low field doublet at 7.55T indicated the presence of 5-methyl-

2-hexynoic acid, and a single peak at 8.70-r corresponded exac tly to that in an authentic sample of lithium pivalate dissolved in n2o.

7. Me tall.a tiPJl• ·.!f Dimethyl Sulfoxide

A volume of 1.5 ml . of a hexane solution containing 3.34 meq .

(1.67 mmo les) of di-s-butylmagnesium (and a trace of DME ) was added with shaking and cooling (ice-bath) to a solution of 0.25 ml . (3.4 mmo les) of dimethyl sul foxide (DMSO) in 3 ml . of benzene contained in a nitrogen-flushed centrifuge tube fitted with a rubber serum cap.

Re.ac tion was vigorous (gas evolution, formation of a precipitate) throughout the addition. After shaking the tube vigorously for a few minutes , 0.5 ml . (5 mmo les) of itt.ime·thylchlorosilane was added to the centrifuge tube. After vigorous shaking a fine white preci pitate slowly formed and the mixture was centrifuged. Analysis of the clear supernatant solution by VPC and NMR indicated that there was present in the gas being evolved considerable dissolved a-butane, but only traces of DMS and DMSO were present , and that the trimethyl silylated derivative of DMSO was present . VPC and NMR characteristics of the latter were identical to those of thi s derivative produced on reaction of trimethylchlorosilane with authentic "Dimsyllithium, "

LiCH2SOCH3 , supplied by Lithium Corporation of America. 148

I. ATTEMPTED HYDROGENOLYSIS OF DIALKYLMAGNESIUM REAGENTS

1. With and Without Added TMEDA

Following the procedure and apparatus for hydrogenolysis of

100 alkyllithium reagents developed by Screttas , b 13 ml . of 0.77N

di-�-butylmagnes ium (5 mmo les) in hexane-cyclohexane and 37 ml . of

�-hexane were mixed together and cooled to 0°C in the apparatus under vacuum. After filling the system to slightly above atmo spheric pres-

sure with hydrogen gas , stirring was begun and any pres sure changes noted. No hydrogen ab sorption occurred after 6 min. , and 0.1 mole of

TMEDA per mo le of di -�-butylmagnes ium was inj ected into the system. No hydrogen uptake occurred after another 6 minutes and another 0. 1 mole of TME DA per mole of (�-C4H9) 2Mg was added. Still no hydrogen uptake was observed and none was taken up as more TMEDA was added. Period-

ical ly additional quantities of TMEDA {up to a total of 11 .3 mmoles) were added.

2. --In the --- Presence of Potassium =t- Butoxide Alone ---and With ----Added ----- THF

and TMEDA

A weight of 0.112 grams (1 .0 mmo le) of potassium �-butoxide was placed in the hydrogenolysis unit and, after it was purged wi th hydrogen and evacuated, 37 ml . of �-hexane was added with stirring followed by 13 ml . of 0.77N di-�-butylmagnes ium in hexane-cyc lohexane .

A brown sticky solid formed on addition of the (�-C4H9) 2Mg . The pressure was raised to slightly above atmospheric wi th hydrogen gas .

When no hydrogen uptake occurred after 9 minutes , 0.4 ml . (5 mmo les) 149

of THF was added. No uptake occurred after 15 minutes , and another 0.4 ml . of THF was added. After :23 minutes 0. 75 · mL (5 1JDI10les) of TMEDA

was added, and after 30 minutes another 0.75 ml . TMEDA. No hydrogen

uptake occurred at any time during and after these additions . After 36 minutes 1.1 ml . (10 mmo�es) of anisole was added. No color change was

observed as occurred in the metallation of resorcinol ' dimetllyl ether under comparable conditions .

In another similar run, 0.56 g. (5 mmoles) of potassium !-butoxide was reacted with 13 ml . of 0.77N di-�-butylmagnesium (5 mmoles) in the presence of 37 ml . of hexane containing 1.5 ml . (10 mmoles)· of TMEDA.

The mixture turned a deep yellow and then after a few minutes a deep orange-brown with brown particles sticking to the walls of the flask.

Hydrogen was admitted to atmospheric pressure . A total of 0.6 ml . of hydrogen was ab sorbed in 10 minutes (0.3 per cent of theory) , but no further absorption occurred thereafter.

3. 1.!l the Presence of ;!-Butyl lithium and .!. Limited Amount of TMEDA

Five mmoles of di-�-butylmagnesium and 5 mmoles of �-butyllithi�m contained in a total of 3 ml . of hexane wer� mixed with 0.37 ml . (2 .5 mmo les) of TMEDA in 37 ml •. of hexane in the hydrogenolysis unit under vacuum. Hy drogen was passed in to slightly greater than 1 atm. After

30 minutes , the total hydrogen absorption was 0.3 ml .

J . ATTEMPTED ADDITION OF DIALKYLMAGNESIUM REAGENTS TO ETHYLENE

Runs exactly comparable in size (and procedure) to those described in (I) above were carried out using ethylene instead of 150 hydrogen. No uptake of ethylene occurred in the presence of TMEDA alone or in the presenc e of potassium· !�butoxide , alone or in combina- tion with THF or TMEDA. In the presence of ,!-butyllithium and a limited amount of TMEDA , a total of 1.7 ml . of ethylene was ab sorbed (less than

1 per cent of theory) in 50 minutes.

K. REACTIONS OF DIALKYLMAGNESIUM REAGENTS WITH

KETONES (BENZOPHENONE)

1. Di�•-Butylmagnesium and Benzophenone

To 1 ml . of an approximately 1.5N di-!-butylmagnesium solution in benzene was added 2 ml . of a 15 wt. per cent solution of benzophenone

(£!. 1.5 mmoles) in benzene . Slow addition of the first few drops of benzophenone solution caused a transient color formation (brown) in the solution . Adding the remainder of the solution rapidly (giving a molar excess of ketone » organometallic) caused the brown color to persist indefinitely, i.e. , several days. NMR examination of this solution showed the presenc e of benzophenone . In addition, aliphatic proton signals appeared at 8.32< and 8.52< in a ratio of 1:4; these were presumably singlets from !-butyl;�netqyl groups, but the smal ler lower field peak could be a narrow mul tiplet . A relatively broad unresolved region of signals between 8'• .6t and 9. 2< were also present in the original di-!- butylmagnes ium in benzene solution, due to n- pentane impurity. A small rather broadened peak also appeared at 5.15t possibly due to the methine hydrogen' of. a· .benzhydrol salt. The ,!-butyl signal at 8.52T appeared downfield by about 18 .5 cps from its normal position Vith di -t-butylmagnesium in a hydrocarbon solvent . After hydrolysis of a smal l l 151

portion of the react ion mixture , VPC analysis of the organic layer

showed the presence of isobutane .

Addition of more di-t--butylmagnesium solution (ca--- 1.0 ml .,

so that the organometallic reagent was now in excess) to the above

stable brown-colored mixture caused the color to disappear ins tanr ·

taneously . NMR examination of the decolorized solution showed a

change in the ratio of the peaks at 8.3� and 8.52< (from· l:4 to

1:20) , a loss of the benzophenone spectrum , an increase in the size

of the peak at 5.15< and a profile change in the broad upfield region .

Hydrolysis of th is solution and NMR examination of the upper layer

showed the two peaks at 8.3< and 8.5T (now at a ratio of 1:6.5), a

noticeable shoulder on the benzene peak , and a considerably increased

and clearer region between 8.7< and 9.2<. This region now shows three

distinct peaks with the upfield two most probably representing the isobutane doublet normally found at �· · 9.1<. The lower of these three

peaks at �· 8.8T could be due to the presence of !-butanol , 8.7&r .

An ESR spectrum was obtained from another colored product solu-

tion from the reaction of equal volume s of 15 per cent benzophenone in

benzene and the 1.5N di-!-butylmagnesium solution. A 44 line spectrum was obtained. Both the color and the ESR spectrum deteriorated after

about one hour.

2 . Di-A-Butylmagne sium and Benzophenone

a. NMR studies � the colored complexes . To 0.25 ml . of a 30 wt .

per cent solution (supersaturated) of benzophenone (�. 0.39 mmoles) in 152 cyclohexane in an NMR sample tube there was added in one portion 0�5

ml . of 0.77N di-�-butylmagnesium (£!. 0.2 mmoles) in hexane-cyclohexane.

·A green, sticky , curdy precipitate formed immediately after the forma-

tion of an initial transient pink color . It was impossible to centri

fuge this mixture in order to obtain a spectrum. Another 0.5 ml . of

the di-�-butylmagnesium solution was added to the tube and the green

color disappeared; the precipitate became fine and particulate and

remained suspended in the solution. One half of this mixture was

transferred by syringe to another NMR tube and both tubes centrifuged.

NMR examination of the clear supernatant solution in one tube

showed little or no di-�-butylmagnesium , but did show signals in the

aromatic region just upfield {10-20 ctp�} from the benzene peak; possib ly

that signal originated from the Mg salt of benzhydrol . Addition of 0.5

ml . of dry benzene to this NMR tube caused solution of most of the solid.

The NMR spectrum of this solution was not significantly different from

that prior to ad'd::tt:ion of benzene . Another 0. 3-ml . portion of di-s

butylmagnes ium solution was added to the tube, but the �-butyl methine

sextet of lines normally seen with (!-C4H9) 2Mg in the 9-9.5t region did not appear, even at a spectrum amplitude of 1 20.

To the other NMR tube was added 0.5 ml . of ethyl ether; little

or no dissolution of solids occurred. A spectrum of the supernatant

solution now showed a triplet of lines in the 8.9-9.lt region . (This

spectrum has not yet been fully rationalized.)

Solutions with different molar ratios of benzophenone (as a 15

per cent solution in benzene) to di-�-butylmagnesium were examined by 153

NMR . A green solution was obtained at a 1:1 ratio and higher, while at ratios from· 1:1 to 1:2 (ketone : organometallic) , a light violet-pink color was ob tained. With less ketone , the color was no longer observ able. No NMR signal for the methine sextet (9.0- 9.5T)· of di-�-butyl magnesium could be ob tained until the 1:2 ratio (excess organometallic) was attained. ESR spectra of both the green and violet solutions were obtained.

b. Study of the products 2f reaction. To 2 ml . of the 15 wt .

per cent benzophenone solution in benzene (ca . 1.5 mmoles) in a centri

fuge tube fitted wi th a rubber septum was slowly added 4 ml . of 0.77N di-�-butylmagnesium in hexane-cyclohexane (1.5 mmoles) . An intense green color was formed on mixing and heat was generated. (In another experiment it was observed that if small increments of di-�-butyl magnesium solution were allowed to layer on top of the benzophenone solution, a deep violet color resulted in the upper layer; mixing of the layers caused this color to immediately change over to green .) A small additional portion of di -�-butylmagnesium solution was added, £!·

0.5 ml ., bringing the organometallic a mo lar excess over the ketone .

With the addition of this portion, the green color disappeared; no heat was generated; no violet color formed; precipitation of solids occurred. The product mixture in the centrifuge tube was poured into a mixture of ice and 3N aqueous HCl to effect hydrolysis . The mixture was shaken well, the layers separated and the aqueous layer was washed with ether. The combined organic layers were treated with anhydrous 154

Na2C03 and evaporated to dryness on a steam bath . The residue was

heated for a few minutes on a hot plate and dissolved in CCl4. NMR

· examination indicated the presence of benzhydrol (CH at 4.53T }� and

the expected addition product (peaks in aliphatic region - 8.3-9.4T)

in a ratio of 2.8: 1 (74 per cent benzhydrol and 26 per cent addition

product) . Some small additional unidentified single peaks appeared

at 7. 20'ral d 7. 55T.

'. L. ATTEMPTED REACTION OF DIALKYLMAGNE SIUM REAGENT WITH

PHENYLDIMETHYLCHLOROSILANE

To 1.0 ml . of a 1.3N solution of di-�-butylmagnesium in ether

there was added 2.5 ml . of phenyldimethylchlorosilane (a large molar

excess) . No apparent reaction resulted, as indicated by a lack of

precipitate formation and no heat of reaction . The reaction mixture

was distilled until its b. p. reached 90°. Ro. change in:�appearance: ·was

npted ;::no · saH ptecip1tated .-·:::The· mixture was. discarded .

M. EFFECT OF DIALKYLMAGNESIUM REAGENTS ON THE HALOGEN.-�TAL

INTERCONVERSION REACTION WITH ALKYLLITHIUM REAGENTS

It was noted that only a slight �itive test (light orange color)

was obtained with Gilman Color Test IIa on the supernatant solution from

reaction of 100 ml . of a 1.3N !,-butyllithium in cyc lohexane solution with

50 ml . of a MgCl2 slurry in cyclohexane . (The MgC 1 2 had been prepared

by reaction of chlorine gas with �-butylmagnesium chloride in ether.) A

Mg analysis of this supernatant solution showed that only about 50 per 155 cent exchange of Mg for Li had occurred. The solution was found to be 0.4N in �-butyllithium, wh ich co�entration alone gives a strong

Gilman Color Test IIa, intense red color.

To 3.5 mL of the above supernatant solution, containing 0.4N lithium reagent and 0.4N magnesium reagent , in a centrifuge tube fitted wi th a rubber septum was added 5 ml . of a 15 wt . per cent

�-bromodimethylaniline solution in benzene. To 3.5 ml. of 0.4N

�-butyllithium (4.5 mmoles) in cyclohexane contained in another capped centrifuge tube was added 5 ml . of the- �-bromodimethylaniline solution. After mixing the reagents in each tube for about one minute , the solution was withdrawn wi th a syringe and carbonated by addition to a stirred Dry Ice-ether slurry.' After the mixtures from each tube had come to room temperature, the ether was evaporated on a steam bath and the residual salts were transferred to centrifuge tubes and washed once wi th ether . The salts were blown dry under a nierogen stream at 50-60°C (oil· bath) . On cooling, the salts were di ssolved in 020 and the tubes centrifuged. NMR examination yielded the follow ing results . From the �-butyllithium solution: no salt of 2-methyl butyric acid was present; two doublets of equal intensity were found in the aromatic region as well as a singlet above in the aliphatic region due to the six protons of the dimethy lamino group ; integration of the aromatic protons and the dimethylamino protons gave the correct ratio (4:6) for lithium �- dimethylaminobenzoate. From the 2: 1 �-butyl lithium- di -�-butylmagnesium solution: 2-methylbutyric acid salt was found to be the main product, at least 95 per cent of the total ; only a trace of the salt of �- dimethylaminobenzoic acid was present. 156

N. HOMOLOGIZATION OF ALKYLLITHIUM REAGENTS AND DIMETHYL ETHER

1. &-Butyl lithium

· a. Determination of pr oducts of reaction. Into 50:-ml. . of:; a

1.24N �-butyllithium solution in- n-hexane (0 .062 mole) , was passed

(with ice bath cool ing) approximately 5 grams of dimethyl ether (0 .10 mo le) and the mixture was allowed to warm to room temperature with

stirring . After two hours VPC analysis of a hydrolyzed sample of the _ solution indicated that there was in the solution a DME :2-methyl

butyllithium: �-butyllithium ratio of 8:4: 1. The 2-me thylbutane , which

identified the 2-methylbutyllithium, was obtained by hydrolysis of an

authentic sample of the lithium reagent prepared from·lit hium metal and

2-methyl -n-butyl bromide in pentane. Two columns were used to insure ° ° positive identification (SE-30 at 33 C, and DEG-AgN03 at 25 C) . The NMR spectrum of the homolo.gizat ion·:product� waa, found to be .. identical 'with

that of the authent ic 2-methylbutyl reagent . A trimethylsilyl deriva

tive from a similar run , one which contained MgCl 2 , was prepared by reac tion of a sample wi th excess trimethylchlorosilane ; the derivative was found to have the same VPC retention time (SE-30 column at 51°)

as an authentic sample prepared from 2-methylbutyllithium and excess

trimethylchlorosilane .

b •.St udy of the kinetics of the reaction.

(1) Uncatalyzed. Into a 200 ml ., 3-necked, round-bottom flask equipped with a magnetic stirring bar, reflux condenser, and thermometer was placed 100 ml . of 1.35N �-butyllithium in cyclohexane . The contents of the flask were cooled to 0-2°C. 157 A lectu.re bottle of dimethyl ether was weighed on a bialance and

then attached by a length of rubber tubing to a gas inlet tube inserted

in one neck of the flask. The lecture bottle was weighed again and

the needle .valve barely opened and DME passed into the stirred solu

tion until the lecture bottle had decreased in weight by about 9 to 9.5

g. (about 10 minutes) . The temperature during the addition was �in-

· tained between 2 and 8°C (ice bath) . The rubber tubing was then

removed, the inlet tube replaced with a rubber septum, and the lecture

bottle weighed again; 9.5 g. of DME were used (100 per cent in excess

of stoichiemetry) . A sample of the solution was removed for NMR anal

ysis at this point, NMR· analysis of the solution 2 minutes after sample

removal showed the methylene doublet of 2-methylbutyllithium emerging

from the me thine sexte.t of ,!-butyllithium. Within 5 minutes after

sample removal, it was already difficult to pick out the methine

sextet. The ice bath was removed and the temperature of the solution

brought as quickly as possible to 40°C by means of a pre-heated oil

bath and ke�t between 38-42°C throughout the study. One mi lliliter

samples of the mixture were removed periodically (at 15 min. , 45 min. ,

1 hr ., 1.5 hr ., 7.0 hr . , and 23 hr .) wi th a syringe and added to 15 mL, _

centrifuge tubes fitted with rubber septa and containing 1-ml . quanti

ties of trimethylchlorisilane . After reac tion was complete (no further

precipitation· of LiCl) , the tubes were centrifuged and the clear super

natant solutions analyzed for trimethylsilyl derivatives of ,!- butyl

and 2-methylbutyllithium by VPC on an 8 ft . , 1/8 inch SE-30 column at

7 5°C (helium carrier gas , flow 7) . Area counts were obtained by , 158 mu ltiplying peak heights of the two derivatives by their distances

from the air peak (33 and 55 .7 units respectively) . The reaction �as

at least 95 per cent complete. after J hours . After 23 hours , the clear

supernatant solution had a total base concentration of 0.39N. · . VPC .

showed that the solid residue contained lithium me thoxide.

The above run was repeated at a lower temperature (5°C) . · simi

lar runs were also made' at 0° and at 40°C using just a stoichiometric

amount of dime thyl ether , i.e. , such that the molar ratio of �-butyl

lithium to DME was 2 :1.5.

(2) Catalyzed. After obtaining a series of kinetic results as

de scribed above , incremental amounts of TMEDA were added to runs with

stoich iometric DME and new series of kinetic data generated. At 0°

the concentrations of TMEDA used were 0.010 and 0.033 moles per liter .

At 40° the concentration of 1�DA was 0.024 M. A run with 100 per

cent excess DME (as described above) and containing 0.013 mQles r.MEDA

per liter of solution was carried out at 2.5° .

2 . ��Butyl lithium

A volume of 82 ml . of a 1.75N solution of S-butyllithium in

. �- pentane (0.192 mo les) was cooled to -12°C in a Dry Ice-acetone bath

and 9.1 g. (0 . 198 moles) of DME (37 per cent excess DME over stoi

chiometry) was added as previously de scribed. The temperature was

kept between -6 and -7°C throughout the run . On completion of the

addition, the mixture was bright yellow. (An ESR spectrum was

obtained from this yellow solution and a weak but de finite signal was obtained from a paramagnetic species present.) The syringe had 159

to be pre-cooled wi th Dry Ice to take samples for reaction with trt.�

methylchlorosilane , because of the high volatility of the solution,

otherwise analyses were done as with .n-butyl runs . After 2 hours

reaction was 95 per cent complete and the yellow color was fading

out ; reaction was continued to over 99 per cent completion (almost

6 hours) and the yellow color was gone .

In addition to VPC characterization of the trimethylsilyla�

tion product from the homoloziation reaction, a carbonation product

was prepared. The reaction mixture was poured into a stirred Dry··Ice

ether slurry and the resulting salt slurry evaporated to dryness on. ·a

steam bath. The salts were washed 3 times wi th ether on a Buchner

funnel, dried on a steam bath and a portion of them (0. 5300 g.) di s

solved in 1 ml . of D20 . M� thanol (100 pl) was added as an internal

standard. NMR examination showed the two peaks besides those for

me thanol and hydroxyl , characteristic of 3,3-dimethylbutyric acid�

0. GENERAL VPC ANALYSIS TECHNIQUE

For the preparat ion of organometallic solutions for VPC analysis ,

0.5 to 1.0 ml . of the solution was syringed slowly into a cooled 15

ml . conical test tube fitted with a rubber septum and containing

£!· 1 ml . of isopropyl or isoamyl· alcohol . The tube was shaken

thoroughly and several drops of 6N HC l added to clarify the solution .

The tube was centrifuged and 3-5 pl of the clear solution were syringed

into the chroma tograph for analysis. A similar technique was employed

to determine the amount of ether or alcohol remaining in a desolvated 160 slurry of anhydrous MgC lz in hydrocarbon media . In this case 1 ml . of the slurry was syringed into a conical test tube containing an equal vo lume of an alcohol . Af ter thorough shaking, the tube was centrifuged and the clear supernatant drawn off for VPC analysis . CHAPTER IV

SUMMARY

Dialkylmagnes ium reagents have been prepared by techniques involving (1) direct reaction of magnesium metal with alkyl halides in hydrocarbon media containing a limited amount of a.

Lewi s base such as dime thyl ether, (2) reaction of Grignard reagents with stoichiometric amounts of lithium alkyls, and (3) reaction of

"activated" magnesium chloride with lithium alkyls in hydrocarbon media . Those dialkylmagne sium reagents prepared in media contain ing ethers could be desolvated of the Lewis base by continuous codis tillation with hydrocarbon solvents . It has also been found that alkylmagnesium chlor ides in ethereal media can be desolvated in a simi lar manner and thereby converted to a mixture of insoluble magnesium ch lor ide and soluble dialkylmagnesium reagents.

Reagent (RzMg) , with long chain g- alkyl groups or with a- and

B-branched groups (e .g. , �-butyl' or neopentyl) are soluble in hydro carbon solvents. Other reagents (R' zMg ) are relatively insoluble in hydrocarbons , but di ssolve if the hydrocarbon contains one mo le of

Lewis base per gram-atom of magne sium. Admixture of a soluble reagent ,

RzMg , with an insolub le one , R'zMg , solubilize& the latter in hydro carbons . Admixture of R'zMg with a hydrocarbon solution of a lithium alkyl also causes dissolution. These solubility properties indicate the formation of mixed reagents, e.g. , RMgR ' and R3MgLi , with specific structural features .

161 1 62

Solid complexes of dialkylmagnesium compounds were formed with triethylenediamine and with methyltriethylenediamine , R'2Mg forming

1:1 and 2:1 complexes and the branched R2Mg forming only 1:1 complexes . NMR analyses of the dialkylmagnes ium compounds and of their complexes confirmed the existence of mixed reagents with specific struc tural features. In the presence of dialkylmagnes ium reagents the normally facile halogen-metal interconversion reac tion between £-bromodimethyl aniline and a lithium alky l was suppressed almost completely, also indicating the formation of intermetallic reagents between lithium and magnes ium alkyls.

Successful me tallations of several different aromatic substrates,

ArH , were carried out with R2Mg reagents in the absence of Lewis bases; in some cases a mixed reagent� ArMgR, could be isolated. These metal lations and certain other reactions did not occur with Lewis bases present . Such results are not consistent with interpretation of magnes ium reagent reactivity in the simple terms of carbanionic character.

NMR and ESR studies of the addition of dialkylmagnes ium reagents to benzophenone in hydrocarbon solvents showed that under certain condi tions and at certain preferred reagent ratios , stab le paramagnetic , colored complexes could be formed. Tentatively , these complexes are represented as containing radical anions wh ich arise both by the addi tion of one mo le of R2Mg to benzophenone , followed by transfer of an electron to excess benzophenone to form a charge-transfer complex

(4-1), and by the direct formation of a charge transfer species from 163 ben�ophenone in the presence of excess R2MS (4-2). 1 ' .! ROMgR + Ph2CO + [ROMgR ]+ (PhiCO] '(4rl) . + � 7 R2Mg + Ph2CO [R2Mg ] [Ph2 co} (4�2)

It was found in the course of these investigations that �-

�nd s-alkyllithium reagents were homologized to the primary reagent by reaction with dimethy l ether . The kinetics 'of the rea� tion wa s studied and appeared to follow the stoichiometry of the reac tion represented below (4 -3) i.e. , the reaction was found to

2 (RLi) i :DME+ DME + (RCH2Li) 2:DMEt 2Li0Me + 2RH (4-3) be third order overall in the presence of the indicated stpichiometric amount of DME and pseudo-second order overall in �the presence of sufficient excess of DME . In the presence of catalytic amounts of tetra- methylenediamine (TMEDA) the homologization was found to be first order in lithium reagent and to vary with THEDA concentration. These resul ts are cons istent with the concept of a kinetic role for base ,

DME, or THEDA , being the catalys is of charge trans fer from reagent to substrate ; in (4-3) one solvated dimer acts as reagent , · the other �s substrate and DME is a catalyst- THEDA acts as such a good charge- transfer catalyst that it washes out the order with respect to the reagent. During reaction, the yellow S-butyllithium�DME mixture was found to be paramagnetic . BIBLIOGRAPHY BI:SLIOGRAPHY

L H. Gilman and R. E. Brown , �· Trav. �· , 48 , 1133 (1929) .

2 . Ph •. von L�hr, Ann. , 261 , 72 (1891) .

3. Fleck, Ann . , 276, 138.

4. W. Hallwachs and A. Schafarik, Ann ., �' 206 (1859) .

5. A. Cahours, Ann. , 114 , 240 (1860) .

6. W. Schlenk and w. Schlenk, Jr . , Ber., 62B, 920-4 (1929) .

7. M. S. Kharasch and 0. Reinmuth, "Grignard .Reactions of Nonmetallic Su.bstances," Prentice-Hall, Inc . , N. Y. , 1954 .

8. B. J. Wakefield, Organometallic Chemistry Reviews , 1, (2), pp . 131-157 (1966) .

9� E. C. Ashby , Qu art. Rev., 21, 259 (1967) .

10. R. E. Dessy and W. Kitching in , "Advances in Organometallic Chemistry," Academic Press, N. Y. , 1966, Vol. 4, pp . 280-289 .

11. R. Kullmann , Compt . rend. , 231 , 866-8 (1950) .

12 . W. Strohmeier , Chem . Ber., 88 , 1218-23 (1955) ; W. Strohmeier and F. Seifert, Ber., 94 ,n56 (19 61) .

13. W. Strohmeier and F. Seifert , Ber., �' 2356 (1961) .

14 . A. Kirrmann , M. Vallino , and J. Fauvarque, Bull. Soc . Chim. France, 1963 , (7) , 1408-10.

15. · F. E. Evans and R. C. Huston, �· Org. Chem. , 24, 1178-83 (1959) .

16 . H. Gilman and R. Brown, Rec . Trav . Chim. , 49, 202-4 (1930) .

17. D. 0. Cowan and H. S. Mo sher, �· Org. Chem. , lr• 1-5 (1962).

18. H. S. Mo sher and P. K. Loeffler ,�· Am. Chem. Soc., 78; 4959 (1956) .

19 . W. H. Glaze and C. M. Selman, �· Organometallic Chem. , 1, 477-480 (1966) . 20. V. Tschelinzeff, Ber., li• 4534-40 (1904) .

21. w, Schlenk, Ber. , 64B , 739-43 (1931) .

165 166

22 . L. I. Zakharkin , 0. Yu Okhlobystin and B. N. Strunin , Doklady �· Nauk §§§.!, 147 , 108-110 (1962).

23. D. Bryce-Smith and G. F. Cox, l.· .9!!!!· Soc. , 1175 (1961) .

24. C. Randaccio , !tal. 548183, Septemb er 19 , 1956.

25. S. M. Blitzer and T. H. Pearson , U. s. 2 ,959,625, November 8, 1960.

26. C. M. Suter and H. L. Gerhart, .J.• �� Chem� Soc• , 57, 107-9 (1935).

27. H. Gilman and R. Jones , J. Am. Chem • ..§.2.£. , 63 , 1439 (1941) .

28. K. Ziegler and E. Ho lzkamp , �. , 605 , 93-7 (1957) .

29. P, Kobetz and R. C. Pinkerton , U. S. 3,028,319 , Ap ril 3, 1962 .

30. G. C. Rob inson , U. s. 3,057,894, October 9, 1962 .

31. H. Gilman , "Organic Chemistry, An Advanced Treatise," John . Wiley , N. Y. , Second Ed . 1958 , Vol. 1, p. 490.

32. H. Gilman and J. Swiss, J. �· Chem . Soc� , 62 , 1847-9 (1940) .

33. K. Ziegler and F. Dersch, �. , 64 , 448 (1931) .

•. 34 w. H. Glaze, C. M. Selman , and C. H. Freeman , -----Chem. f2!!., 14, 474 (1966) .

35. J. F. Nobis and R. E. Robinson , U. S. 2 ,795,625, June 11, 1957.

36. W. N. Smith , Jr . , "Studies of Organometallic Compounds," Doctoral Dissertation , The University of California, Berkeley , California, 1963.

37. A. Tzschach and H. Hackert, �· Chem. , iQl, 266 (1966).

38. E. Weiss, J.,Organometal. fh!!., 1. 314-321 (1964) .

39. E. Weiss, Chem. Ber. , 98, 2805-2813 (1965) .

40. A. ·Kirrmann , R. Hamelin , and S. Hayes , !!!!!.· Soc� £!!.!!!. France, 1395-1403 (1963) .

41. S. J. Storfer and E. I. Becker, .d.· Org� £h!!., 12. 1868 (1962).

42. E. C. Ashby , .d.· Am• .9!!!!· Soc. , .[l., 2509 (1965) .

43. A. D. Vreugdenhil and C. Blomberg , �· .I!.!Y.· Chim. , 82 , 453 (1963) . -'

167

44 . E. C. Ashby and M. B. Smith, !· .Am. fh!m. Soc., 86 , 4363 (1964) . 45 . M. B. Smith and W. E. Becker, Tetrahedron Le tters, 43 , 3843 . (1965) .

46 . E. C. Ashby and F. W.alker, !· Organom.e tal . fh!m. , 2, . 17 (1967).

47 .' A. Scala and E. Becker, !· Org. �. , 30 , 3491 (1965) .

48 . H. Normant , Advan. Org. fh!m. , 1, 6 (1960) .

49 . P. Vink , C. Blomberg , A. D. Vreugdenhil and F. Bickelhaupt, Tetrahedron Letters, 52 , 6419 (1966) ,

50. G. D. Stucky and R. E. Rundle ,!· .Am . Chem. �. , 85, 1002 (1963) .

51. G. E. Coates and J. A. Heslop, � Chem. Soc. , (� 1966, 26.

52 . L. I. Zakharkin,. Zh . Ob shchei Khimi i, 34 (9) ,. 3125 (1964) .

53. G. E. Coates and D. Ridley , :[. Chem. �. , (A) , 1967, 5£i .

54 . G. F. Wr ight , :[. Am. fh!m. Soc., 61 , 1152 (1939) .

55 . A. D. Vreugdenhil and C .. Blomberg, �· �· Chim. , 83, 1096 , (1964) .

56. R. E. Deasy and R. M. Jones , !· Org. Chem. , 24, 1685 (1959) .

" 57. W. Strohmeier , F. Seifert and H. Lande feld, Zeitschrift � Elektrochemie, 66 , 312 (1962).

. 58 . W. St rohmeier a�d F. Seifert , Zeitschrift fUr .Elektro chemie ,63�683(1959) • . . -- . . - 59 . H. F. Ebel and R. Schneider , Angew. Chemie. Int'l. Ed. , �, 878. ;(1965) .

60 . T. Psarras and R. Deasy , !· �· Chem. Soc. , 88, 5132 (1966) .

61. G. Fraenkel , D. Adams , and J. Williams , Tetrahedron Letters , 12 , 767 (1963) .

62. G. Fraenkel and .D. Dix, :[. Am . fh!m. Soc., §..§.., 979 (1966) .

63. G. Whi tesides, F. Kaplan, and J. D. Roberts, !· Am . �· �. , 85 , 2167 (1963) .

64 . M. Witanowski and J. D. Roberts, :[. Am. �· Soc. , 88 , 737 (1966) . 168

65. L. M. Seitz and T. L; Brown , d· Am. Chem. �� ' 88 , 4140 (1966) .

66. R. M. Salinger and H. S. Mosher , .:!· Am. Chem. �· , 86, 1782 (1964) .

• 67. ·T. Holm, Tetrahedron Letters , -28, 3329 (1966) . · .

68. M. Ante1,1nis , .:!· Orp;. Ch�m. , J:l._, 596 (1962).

69. H. 0. House and W. L. Respess ,.:!· Org� �. , 30 , 301 (1965) .

70. G. E• Coates and D. Ridley, Chem. Comm. , 16 , 560 (1966) .

71. D. Bryce-Smith and-I. F. Graham, Chern. Comm. , 16, 559 (1966) . -- -

7 2. H. ·0. House and -D. D. Traficante, .:!• Org. �· , 28, 355 (1963) .

73. c. A. Guthrie, E. ·Y . Spencer, and G. F. Wright , Can • .:!· �· , ]2, 873 (1957) .

74. s. G. Smith, Tetrahedron Letters , I, 409 (1963) .

75 . S. G. Smith and G. Su, .:!· Am. �· �. , 86 , 2750 (1964) .

76. J. H. Wotiz, C. A. Hollingsworth and R. E. Deasy, d· Org. Chem. 21, 1063 (1956) .

77 • J. H. Wotiz , C. A. Hollingsworth , R. E. Deasy , and L. C. Lin, .:!• Org. Chem. , n_, 228 (1958�

78. H. 0. House and · H. W. Thompson , .:!· .Q.!:g_. Chem. , 28, 348 (1963) .

79. S. Jacobsen , V. Bitsch, and J. Munch-Petersen , Acta Chemica Scandinavica, 12, 825 (1963) .

80. H. Gilman and .R. Brown, .:!· Am. �· Soc. ,�' 1181-5 (1930) .

81. J. D. Citron and E. I. Becker , Can. d· �. , 41, 1260 (1963) .

82 . R. E. Deasy and R. M. Salinger, .:!· Am. Chem. Soc. , 83, 3530 (19 61) .

83. F. H. ·Norton and H. B. Hass , .:!· �· Chem. Soc.. , 2.§., 2147 (1936) .

84. C. Goulombic and D. L. Cottle, .:!· Am. C}lem. Soc. ; 61 , 996 (1939) .

85 . R. w. Freedman and E. I. Becker, .:!· Q!&· Chem. , 1&, 1701 (1951) .

86. W. Rel;!ves .and L. w. Fine, .:!· Am. Chem. �. , 86 , 880 (1964) . 169

87 . H. Gilman , E. St. John, N. St. John , and M. Lichtenwalter, Rec . !!!Y· �· ' �. 577 (1936) .

88. L. Lardicci, L. Lucarine , P. Palagi , and P. Pino , J. Organo metallic Chemistry, _i, 341 (1965) .

89 . G. Coleman and R. Blomquist, I· Am. �· �· · 63 , 1692 (1941) .

90. L. I. Zakharkin , 0. Yu Okhlobystin, and K. A. Bilevitch , J. Organo-. metal . Chem. , b 309 (1964) .

91. L. I. Zakharkin , I· Organometal. Chem. , _i, 349 (1965) .

92 . R. Bauer , Zeit. furII Naturforschung, 17B , 201 (1962) .

93 . H. E. Podall and W. E. Foster , I· Org. Chem. , 23, 1848 (1958) .

94. S. K. Podder , E . W. Smalley, and C. A. Hollingsworth, l·. Org. Chem. , �. 1435 (1963) .

95 . R. E. Deasy , F. Kaplan, G. Coe , and R. Salinger , J. Am. Chem . Soc., 85 , 1191 (1963) ; 88, 460 (1966) .

96 . R. Salinger and R. E. Dessy , Tetrahedron Letters, 11, 729 (1963) .

97. A. W. Langer , French Patent 1,425,276 (Dec . 13, 1965) .

98 . (a) M. Fontanille and P. Sigwalt, Compt . Rend . Ser . f_. , 263 , 624 (1966) . (b) s. C. Watson and J. F. Eastham , I· Organometal .Chem. , 2_, 165 (1967) .

99. (a) A. Mendel, I· Organometal �· , i• 97 (1966) .

• 2 (b) R. Oda and K. Yamamoto, I· Org. Chem ., 6, 4679 (1961) . 100 . (a) Arthur Pitts Werts III , �'Applications of Dimethyl Sulfoxide in Organic Synthesis ," Master 's Thesis, The University of Tennessee, Knoxville, Tennessee , 1963 . (b) C. G. Screttas , "Base Catalysis of Formation and Lysis of the Lithium-Carbon Bond," Doctoral Dissertation , .The University of Tennessee, Knoxville , Tennessee, 1966.

101 . P. Pfeiffer and H. Blank , I· Prakt . �· (2 ), 15�-242 (1939) .

102 . G. Ru ssell, E. Janzen , and E. Stromm , I· Am. Chem. Soc., 86, 1807 (1964) .

103 . F. Runge, E. Taeger, C. Fi�dler, and E. Kahlert , J. Prakt . Chem. , 19 , 37 (1963) .

104. L. Selvidge, "Base Catalysis of Hydrogenolysis and Addition Reactions of Lithium Alkyls ," Master 's Thesis , The University of Tennessee, Knoxville, Tennessee , 1967 . 170

lOS. F. Settle , Jr. , "The Structure and Properties of Organolithium Compounds ," Doctoral Dissertation , The University of Tennessee , Knoxville, Tennessee , 1964.

106. C. G. Screttas and J, F. Eastham , .:!• �· fb.!!!· Soc. , 88 , 5668 (1966).

107. J. F. Eastham, .:!· �· Chem. �· · �. 1683 (1957) .

108. A. K. Ts iomis , "A Study of the Complexation Organoaluminum Compounds Using the Conductometric Titrimetry ," Doctoral Dissertation , The University of Tennessee, Knoxville , Tennessee. 1967. .

109. G. G. Barbaras , C. Dillard , A. E. Finholt , T. Wartik , K. Wilzbach ,

and H. Schlesing er, .:!· �· Chem. 12£• , 1.1. 4585 (1951) .

110. H. Zimmer and H. Gold, Ber. , 89 1 712 (1956) ; H. Zimmer and D. George, �· , �� 2285 (1956):-

111. The author appreciates assistance from Dr . J. F. Eastham in mathematical interpretation of the kinetic data. VITA

Conrad William Kamienski was born in New York City on December

23, 1929. He obtained his elementary and secondary education . in New

York City and graduated with a B.S. degree from the City College of New

· York in 1950. After a year of post-gradua te study at Fordham University

in New York , he moved to Albany , New York where he worked for one year as a medical technician. In 1952 , he joined the General Electric Com· pany's Research Labs in Schenectady, New York as a chemical technician where he was employed for five years. He ob tained an M. S. degree from

the Reneeela�r . Polytechnic Institute in 1957 through the benefits of

the General Electric Education and Charitable Fund. He joined the Lithium

Corporation of America in the summer of 1957 as a research chemist and

- is presently employed by this company as its director of organic chemi-

cal research.

In 1965, through the generosity and foresight of the President

of Lithium Corpora tion of America , Mr. Harry D. Feltenstein, he was

permitted to pursue his studies fulltime at the University of Tennessee

toward a Ph .D. degree under the direction of Professor Jerome F. Eastham.

In Decemb er 1967 he was granted the Doc tor of Philosophy degree from

the University of Tennessee .

The author is a member of the American Chemical Society and the

Phi Lamb da Upsilon and Sigma Xi Soc ieties , and the Research Society of

America. �1